%
% \GetFileInfo{gmdoc.sty}
% \title{The \pk{gmdoc.sty} Package\thfileinfo}
% \author{Natror (Grzegorz Murzynowski)}
% \date{\today}
% \maketitle
% \label{gmd}
%%
%% This is (a~documentation of) file \pk{gmdoc.sty},
%% intended to be used with \LaTeXe\ as a~package for
%% documenting \LaTeXpar\ files and to be documented with itself.
%% \stanza
%%
% \begin{copyrnote}
%
%% Written by Natror (Grzegorz Murzynowski),
%% natror at o2 dot pl
%%
%% \copyright\,2006, 2007, 2008 by Natror (Grzegorz Murzynowski).
%%
%% This program is subject to the \LaTeX\ Project Public License.
%% See \url{http://www.ctan.org/tex-archive/help/Catalogue/licenses.lppl.html} ^^A
%% for the details of that license.
%%
%% \acro{LPPL} status: "author-maintained".\par
%%
%% Many thanks to my \TeX\ Guru Marcin Woli\nacute ski for his \TeX nical ^^A
%% support.
%
%\end{copyrnote}
%
% \ChangesStart{v0.98c}{1000/0/0}
%
% \chschange{v0.96}{2006/08/22}{2395}
% \chschange{v0.97}{06/09/04}{3636}
% \chschange{v0.98}{06/09/04}{3636}
% \chschange{v0.98a}{06/09/06}{4092}
% \chschange{v0.98b}{06/9/7}{4119}
% \chschange{v0.98c}{06/9/10}{4299}
% \chschange{v0.98d}{06/9/15}{4409}
% \chschange{v0.98e}{06/9/23}{4420}
% \chschange{v0.98f}{06/9/30}{4430}
% \chschange{v0.98g}{06/10/5}{4294}
% \chschange{v0.98i}{06/10/14}{4352}
% \chschange{v0.98j}{06/10/17}{4372}
% \chschange{v0.98k}{06/10/21}{4416}
% \chschange{v0.98l}{06/10/27}{4401}
% \chschange{v0.98m}{06/11/14}{4349}
% \chschange{v0.99}{06/11/29}{4462}
% \chschange{v0.99a}{2006/12/01}{4479}
% \chschange{v0.99b}{07/01/01}{4479}
% \chschange{v0.99c}{07/3/2}{4486}
% \chschange{v0.99c}{07/3/30}{4475}
% \chschange{v0.99d}{07/4/26}{4555}
% \chschange{v0.99e}{2007/4/29}{4574}
% \chschange{v0.99g}{2007/11/12}{5229}
% \chschange{v0.99i}{2008/8/3}{5247}
% \chschange{v0.99j}{2008/8/3}{5266}
% \chschange{v0.99k}{2008/8/4}{5261}
% \chschange{v0.99l}{2008/8/6}{5225}
% \chschange{v0.99m}{2008/8/9}{5356}
% \chschange{v0.99m}{2008/8/13}{5354}
% \chschange{v0.99n}{2008/8/22}{5409}
% \chschange{v0.99n}{2008/8/30}{5547}
% \chschange{v0.99p}{2008/10/4}{5607}
% \chschange{v0.99q}{2008/10/10}{5603}
% \chschange{v0.99r}{2008/11/22}{5607}
% \toCTAN{v0.99r}{2008/11/22}
%
%
\ifnum\catcode`\@=11 % Why this test here---will come out in chapter
% \gmiflink{The driver}. ^^A~( the very first line
% ^^A and about coming out already! ;-)
\NeedsTeXFormat{LaTeX2e}
\ProvidesPackage{gmdoc}
[2008/11/22 v0.99r a documenting package (GM)]
\fi
%
%
%
% \CharacterTable {In my \TeX\ Guru's opinion the idea of
% checking of the correctness of chars is a~bit obsolete
% nowadays. Therefore I~define the \CharacterTable<#1>
% macro to gobble its argument and to typeout a~request
% for request.}
%
%
% \newcommand*\docfm{\pk{doc}}
% \newcommand*\gmdoc{\pk{gmdoc}}
%
% \tableofcontents
%
%
% \division{Readme}
%
% This package is a~tool for documenting of \LaTeXpar\
% packages, classes etc., i.e., the \file{.sty}, \file{.cls} files
% etc. The author just writes the code and adds the commentary
% preceded with |%| sign (or another properly declared). No special
% environments are necessary.
%
% The package tends to be (optionally) compatible with the standard
% \pk{doc.sty} package,
% i.e., the \file{.dtx} files are also compilable with \pk{gmdoc}
% (they may need very little adjustment, in some rather special
% cases).
%
% The tools are integrated with \pk{hyperref}'s advantages such as
% hyperlinking of index entries, contents entries and
% cross-references.
%
% The package also works with \XeTeX\ (switches automatically).
%
% \begin{gmlonely}
% \subdivision{Installation} Unpack the \file{gmdoc-tds.zip} archive
% (this is an archive conforming the \acro{TDS} standard, see
% \file{CTAN/tds/tds.pdf})% in a~\file{texmf} directory
% or put the \pk{gmdoc.sty}, \pk{gmdocc.cls} and \pk{gmoldcomm.sty}
% somewhere in the \file{texmf/tex/latex} branch on your own.
% (Creating a~\file{texmf/tex/latex/gm} directory may be
% advisable if you consider using other packages written by me. And
% you \emph{have} to use four of them to make \pk{gmdoc} work.)
%
% You should also install \pk{gmverb.sty}, \pk{gmutils.sty} and
% \pk{gmiflink.sty} (e.g., put them into the same \file{gm}
% directory). These packages are available on \acro{CTAN} as separate
% \file{.zip} archives also in \acro{TDS}-compliant \pk{zip}
% archives.
%
% Moreover, you should put the \file{gmglo.ist} file, a MakeIndex
% style for the changes' history, into some \file{texmf/makeindex}
% (sub)directory.
%
% Then you should refresh your \TeX\ distribution's files' database
% most probably.
% \end{gmlonely}
%
%
% \subdivision{Contents of the \pk{gmdoc.zip} archive}
%
% The distribution of the \pk{gmdoc} package consists of the
% following five files and a~\acro{TDS}-compliant archive.
% \begin{verse}
% \pk{gmdoc.sty}\\
% \pk{gmdocc.cls}\\
% \pk{gmglo.ist}\\
% \pk{README}\\
% \pk{gmdoc.pdf}\\
% \pk{gmdoc.tds.zip}
% \end{verse}
%
%
% \begin{gmlonely}
% \subdivision{Compiling of the documentation}
%
% The last of the above files (the \pk{.pdf}, i.e., \emph{this ^^B
% file}) is a~documentation compiled from the \pk{.sty} and
% \pk{.cls} files by running \XeLaTeX\ on the \file{gmdoc.sty}
% twice (|xelatex gmdoc.sty| in the directory you wish the
% documentation to be in, you don't have copy the \file{.sty} file
% there, \TeX\ will find it), then MakeIndex on the \file{gmdoc.idx} and
% \file{gmdoc.glo} files, and then \XeLaTeX\ on \file{gmdoc.sty}
% once more. (Using \LaTeX\ instead of \XeLaTeX\ should do, too.)
%
% MakeIndex shell commands:
%\begin{verbatim}
%makeindex -r gmdoc
%makeindex -r -s gmglo.ist -o gmdoc.gls gmdoc.glo
%\end{verbatim}
% The |-r| switch is to forbid MakeIndex to make implicit ranges since
% the (code line) numbers will be hyperlinks.
%
% Compiling the documentation requires the packages:
% \pk{gmdoc} (\pk{gmdoc.sty} and \pk{gmdocc.cls}), \pk{gmutils.sty},
% \pk{gmverb.sty}, \pk{gmiflink.sty} and also some standard packages:
% \pk{hyperref.sty}, \pk{xcolor.sty}, \pk{geometry.sty},
% \pk{multicol.sty}, \pk{lmodern.sty}, \pk{fontenc.sty} that should
% be installed on your computer by default.
%
% If you had not installed the \pk{mwcls} classes (available on
% \acro{CTAN} and present in \TeX\ Live e.g.), the result of your
% compilation might differ a~bit from the \pk{.pdf} provided in this
% \pk{.zip} archive in formatting: If you had not installed
% \pk{mwcls}, the standard \pk{article.cls} class would be used.
% \end{gmlonely}
%
% \subdivision{Dependencies}
%
% The \pk{gmdoc} bundle depends on some other packages of mine:
% \begin{verse}
% \pk{gmutils.sty},\\
% \pk{gmverb.sty},\\
% \pk{gmiflink.sty}\\
% \pk{gmeometric} (for the driver of The \LaTeXe\ Source)\\
% \end{verse}
% and also on some standard packages:
% \begin{verse}
% \pk{hyperref.sty},\\
% \pk{color.sty},\\
% \pk{geometry.sty},\\
% \pk{multicol.sty},\\
% \pk{lmodern.sty},\\
% \pk{fontenc.sty}\\
% \end{verse}
% that should
% be installed on your computer by default.
%
%^^A\traceon
% % \subdivision{Bonus: \file{base} drivers}
%^^A\traceoff
% As a~bonus and example of \docfm-compatibility
% there are driver files included
% (cf.\ Palestrina, \textit{Missa papae Marcelli} ;-):
%
% \begin{verse}
% \pk{source2e_gmdoc.tex}\\
% \pk{docstrip_gmdoc.tex}\\
% \pk{doc_gmdoc.tex}
%
% \pk{gmoldcomm.sty}\\
% (\pk{gmsource2e.ist} is generated from \pk{source2e_gmdoc.tex})\\
% \end{verse}
%
% These drivers typeset the respective files from the
%\[\hbox{\file{…/texmf-dist/source/latex/base}}\]
% directory of the \TeX Live2007 distribution (they only read that
% directory).
%
% Probably you should redefine the |\BasePath| macro in them
% so that it points that directory on your computer.
%
%
%\division*{Introduction}
%
% There are very sophisticated and effective tools for documenting
% \LaTeX\ macro packages, namely the \pk{doc} package and the
% \pk{ltxdoc} class.
% Why did I~write another documenting package then?
%
% I~like comfort and \pk{doc} is not comfortable enough for me. It
% requires special marking of the macro code to be properly typeset
% when documented. I~want \TeX\ to know `itself' where the code begins
% and ends, without additional marks.
%
% That's the difference. One more difference, more important for the
% people for whom the \docfm's conventions are acceptable, is that
% \gmdoc\ makes use of \pk{hyperref} advantages and makes
% a~hyperlinking index and toc entries and the
% cross-references, too. (The \CSs in the code maybe in the future.)
%
% The rest is striving to level the very high \pk{doc/ltxdoc}'s
% standard, such as (optional) numbering of the codelines and
% authomatic indexing the control sequences e.g.
%
% The \docfm\ package was and still is a~great inspiration for me and
% I~would like this humble package to be considered as a~sort of
% hommage to it\footnote{As Grieg's Piano Concerto is a~hommage to ^^B
% the Schumann's.}. If I~mention copying some code or narrative but do
% not state the source explicitly, I~mean the \docfm\ package's
% documentation (I~have v2.1b dated 2004/02/09).
%
%
%
% \division{The user interface}
%
%
% \subdivision{Used terms}
%
% When I~write of a~\textbf{macro}, I~mean a~macro in \TeXbook's
% meaning, i.e., a~control sequence whose meaning is
% |\(e/g/x)def|ined. By a~\textbf{macro's parameter} I~mean each of
% |#|\<digit>s in its definition. When I~write about ^^A\)
% a~\textbf{macro's argument}, I~mean the value (list of tokens)
% subsituting the corresponding parameter of this macro.
% (These understandings are according ^^A(
% to \TeXbook, I~hope: \TeX\ is a~religion of Book ;-)\,.)
%
% I'll use a~shorthand for `control sequence', \textbf{\CS}.
%
% When I~talk of a~\textbf{declaration}, I~mean a~macro that expands
% to a~certain assignment, such as |\itshape| or
% |\@onlypreamble{|\<\CS>|}|.
%
% Talking of declarations, I'll use the \textbf{\acro{OCSR}} acronym as
% a~shorthand for 'observes/ing common \TeX\ scoping rules'.
%
% By a~\textbf{command} I~mean a~certain abstract visible to the end
% user as a~\CS but consisting possibly of more than one macro. I'll
% talk of a~\textbf{command's argument} also in
% the `sense\:-for\:-the\:-end\:-user', e.g., I'll talk of the
% |\verb| \emph{command's} argument although \emph{the macro}
% |\verb| has no |#|\<digit> in its definition.
%
% The \textbf{code} to be typeset verbatim (and with all the bells
% and whistles) is everything that's not commented out in the source
% file and what is not a~leading space(s).
%
% The \textbf{commentary} or \textbf{narrative} is everything after
% the comment char till the end of a~line. The \textbf{comment char}
% is a~character the |\catcode| of which is 14 usually i.e., when the
% file works; if you don't play with
% the |\catcode|s, it's just the |%|. When the file is documented with
% \gmdoc, such a~char is re|\catcode|d and its r\ocircum le is else: it
% becomes the \textbf{code delimiter}.
%
% A~line containing any \TeX\ code (not commented out) will be called
% a~\textbf{codeline}. A~line that begins with (some leading spaces
% and) a~code delimiter will be called a~\textbf{comment line} or
% \textbf{narration line}.
%
% The \textbf{user} of this package will also be addressed
% as \textbf{you}.\stanza
%
% Not to favour any particular gender (of the amazingly rich variety,
% I~mean, not of the vulgarly simplified two-element set), in this
% documentation I~use alternating pronouns of third person ^^A(
% (\TextUsage\heshe\ etc. commands provided by \pk{gmutils}), so let
% one be not surprised if `\heshe' sees `\himher self' altered in the
% same sentence :-)\,.
%
%
% \subdivision{Preparing of the source file}
%
% When \LaTeXpar\ with \pk{gmdoc.sty} package loaded typesets the
% comment lines, the code delimiter is ommitted. If the comment
% continues a~codeline, the code delimiter is printed. It's done so
% because ending a~\TeX\ code line with a~|%| is just a~concatenation
% with the next line sometimes. Comments longer than one line are
% typeset continuously with the code delimiters ommitted.
%
% The user should just write \hisher\ splendid code and brilliant
% commentary. In the latter \heshe\ may use usual \LaTeXpar\ commands.
% The only requirement is, if an argument is divided in two lines, to
% end such a~dividing line with \TextUsage*{\^^M} (|\|\<line end>) or with
% \TextUsage*{^^B} sequence that'll enter the (active) \<char2> which
% shall gobble the line end. ^^A~and the leading code delimiter of
%^^A % the next line---obsoleted by ^^A~making % ignored in the commentary.
%
% Moreover, if \heshe\ wants to add a~meta-comment i.e., a~text that
% doesn't appear in the code layer nor in the narrative, \heshe\ may
% use the \TextUsage*{^^A} sequence that'll be read by \TeX\ as
% \<char1>, which
% in \gmdoc\ is active and defined to gobble the stuff between itself
% and the line end.
%
% Note that |^^A| behaves much like comment char
% although it's active in fact: it re|\catcode|s the special
% characters including |\|, |{| and |}|
% so you don't have to worry about unbalanced braces or \cs{if}s in
% its scope. But |^^B| doesn't re|\catcode| anything (it would be
% useless in an argument) so any text between |^^B| and line end has
% to be balanced.
%
% However, it may be a~bit confusing
% for someone acquainted with the \docfm\ conventions. If you don't
% fancy the |^^B| special sequence, instead you may restore the
% standard meaning of the line end with the \TextUsage\StraightEOL\
% declaration which
% \acro{OCSR}. As almost all the control
% sequences, it may be used also as an environment, i.e.,
% |\begin{StraightEOL}|\ \dots\ |\end{StraightEOL}|. However, if for any
% reason you don't want to make an environment (a~group), there's
% a~|\StraightEOL|'s counterpart, the \TextUsage\QueerEOL\ declaration that
% restores again the
% \StraightEOL
% queer\footnote{In my understanding
% `queer' and `straight' are not the opposites excluding each
% other but the counterparts that may cooperate in harmony for
% people's good. And, as I~try to show with the \cs{QueerEOL} and
% \cs{StraightEOL} declarations, `queer' may be very useful and
% recommended while `straight' is the standard but not necessarily
% normative. ^^A% (Remember also Alice's in the Wonderland exclamations
% ^^A % ``What a~queer day is today''.)
% }^^A
% \QueerEOL
% \gmdoc's meaning of the line end. It \acro{OCSR}, too. One more point to use
% |\StraightEOL| is where you wish some code lines to be executed
% both while loading the file and during the documentation pass
% (it's analogous to \docfm's not embracing some code lines in
% a~\env{macrocode} environment).
%
% \stanza
%
% As in standard \TeX ing, one gets a~paragraph by a~blank line.
% Such a~line should be |%|ed of course. A~fully blank line is
% considered a~blank \emph{code line} and hence results in
% a~vertical space in the documentation. As in the environments for
% poetry known to me, subsequent blank lines do not increase such
% a~space.
%
% Then \heshe\ should prepare a~main document file,
% a~\textbf{driver} henceforth, to set all the
% required formattings such as |\documentclass|, paper size
% etc., and load this package with a~standard command
% i.e., |\usepackage{gmdoc}|, just as \docfm's documentation says:
%
% \begin{quotation}
% If one is going to document a set of macros with the \pk{[gm]doc}
% package one has to prepare a special driver file which produces
% the formatted document. This driver file has the following
% characteristics: \par\leftskip\CodeIndent
% \stanza
%
% \noindent|\documentclass[|\<options>|]{|\<document-class>|}|\par
% \noindent|\usepackage[|\<options, probably none>|]{gmdoc}|\par
% \<preamble>\par
% \noindent|\begin{document}|\par
% \<special input commands>\par
% \noindent|\end{document}|\par
% \end{quotation}
%
%
% \subdivision{The main input commands}
%
% To typeset a~source file you may use the \TextUsage\DocInput\
% macro that takes the (path and) name of the file \emph{with ^^B
% the extension} as the only argument, e.g.,
% |\DocInput{mybrilliantpackage.sty}|.
%
% (Note that an \emph{installed} package or class file is findable to
% \TeX\ even if you don't specify the path.)
%
% If a~source file is written with rather \docfm\ than \gmdoc\ in
% mind, then the \TextUsage\OldDocInput\ command may be more
% appropriate (e.g., if you break the arguments of commands in the
% commentary in lines). It also takes the file (path and) name as the
% argument.
%
% When using |\OldDocInput|, you have to wrap all the code in
% \TextUsage*{macrocode} environments, which is not necessary when you
% use |\DocInput|. Moreover, with |\OldDocInput| the\
% \env{mac\-ro\-co\-de(*)} environments require to be ended with
% |% \end{macrocode(*)}| as in \docfm. (With |\DocInput| you are
% not obliged to precede |\end{macrocode(*)}| with The
% Four Spaces.)
%
% \stanza
% If you wish to document many files in one document, you are
% provided \TextUsage\DocInclude\ command, analogous to \LaTeX's
% |\include| and very likely to \pk{ltxdoc}'s command of the same
% name. In \gmdoc\ it has one mandatory argument that should be the
% file name \emph{without extension}, just like for |\include|.
%
% The file extensions\label{docinclude:extensions} supported by
% |\DocInclude| are \pk{.fdd}, \pk{.dtx}, \pk{.cls}, \pk{.sty},
% \pk{.tex} and \pk{.fd}. The macro looks for one of those extensions
% in the order just given. If you need to document files of other
% extensions, please let me know and most probably we'll make it
% possible.
%
% |\DocInclude| has also an optional first argument that is intended
% to be the path of the included file with the levels separated by |/|
% (slash) and also ended with a~slash. The path given to |\DocInclude| as
% the first and optional argument will not appear in the headings nor
% in the footers.
%
% |\DocInclude| redefines \TextUsage\maketitle\ so
% that it makes a~chapter heading or, in the classes that don't
% support |\chapter|, a~part heading, in both cases with
% respective toc entries. The default assumption is that all the
% files have the same author(s) so there's no need to print them in
% the file heading. If you wish the authors names to be printed, you
% should write \TextUsage\PrintFilesAuthors\ in the preamble or before
% the relevant |\DocInclude|s. If you wish to undeclare printing the
% authors names, there is \TextUsage\SkipFilesAuthors\ declaration.
%
% Like in \pk{ltxdoc}, the name of an included file appears in the
% footer of each page with date and version info (if they are
% provided).
%
% The |\DocInclude|d files are numbered with the letters, the
% lowercase first, as in \pk{ltxdoc}. Such a~filemarker also precedes
% the index entries, if the (default) codeline index option is in force.
%
% As with |\include|, you may declare
% \TextUsage\includeonly|{|\<filenames separated by commas>|}| for
% the draft versions.
%
% \stanza If you want to put the driver into the same \file{.sty} or
% \file{.cls} file (see chapter \ref{Driver} to see how), you may
% write |\DocInput{\jobname.sty}|, or |\DocInclude{\jobname.sty}|, but
% there's also a~shorthand for the latter \TextUsage\SelfInclude \
% that takes no arguments.
% By the way, to avoid an infinite recursive input of \file{.aux}
% files in the case of self-inclusion an \file{.auxx} file is used
% instead of (main) \file{.aux}.
% \stanza
%
% At the default settings, the |\Doc/SelfInclude|d files constitute
% chapters if |\chapter| is known and parts otherwise. The
% |\maketitle|s of those files result in the respective headings.
%
% If you prefer more \pk{ltxdoc}ish look, in which the files always
% constitute the parts and those parts have a~part's title
% pages with the file name and the files' |\maketitle|s result in
% (article-like) titles not division headings, then you are provided
% the \TextUsage\ltxLookSetup\ declaration (allowed only in the
% preamble). However, even after this declaration the files will be
% included according to \gmdoc's rules not necessarily to the \docfm's
% ones (i.e., with minimal marking necessary at the price of active
% line ends (therefore not allowed between a~command and its argument
% nor inside an argument)).
%
% On the other hand, if you like the look offered by me but you have
% the files prepared for \docfm\ not for \gmdoc, then you
% should declare \TextUsage\olddocIncludes. Unlike the previous one,
% this may be used anywhere, because I~have the account of including both
% \docfm-like and \gmdoc-like files into one document. This
% declaration just changes the internal input command and doesn't
% change the sectioning settings.
%
% It seems possible that you wish to document the `old-doc' files
% first and the `new-doc' ones after, so the above declaration has its
% counterpart, \TextUsage\gmdocIncludes, that may be used anywhere,
% too. Before the respective |\DocInclude|(s), of course.
%
% Both these declarations \acro{OCSR}.
%
% If you wish to document your files as with \pk{ltxdoc} \emph{and}
% as with \docfm, you should declare
% |\ltxLookSetup| in the preamble \emph{and} |\olddocIncludes|.
%
% \stanza
% Talking of analogies with \pk{ltxdoc}, if you like only the page
% layout provided by that class, there is the
% \TextUsage\ltxPageLayout\ declaration (allowed only in preamble)
% that only changes the margins and the text width (it's intended to
% be used with the default paper size). This declaration is
% contained in the |\ltxLookSetup| declaration.
%
% \stanza
% If you need to add something at the beginning of the input of file,
% there's the \TextUsage\AtBegInput\ declaration that takes one
% mandatory argument which is the stuff to be added. This declaration
% is global. It may be used more than one time and the arguments of
% each occurrence of it add up and are put at the beginning of input
% of every subsequent files.
%
% Simili modo, for the end of input, there's the
% \TextUsage\AtEndInput\ declaration, also one-argument, global and
% cumulative.
%
% If you need to add something at the beginning of input of only one
% file, put before the respective input command an
% \TextUsage\AtBegInputOnce|{|\<the stuff to be added>|}|
% declaration. It's also global which means that the groups do not
% limit its scope but it adds its argument only at the first input
% succeding it (the argument gets wrapped in a~macro that's |\relax|ed
% at the first use). |\AtBegInputOnce|s add up, too.
%
% \stanza
% One more input command is \TextUsage\IndexInput\ (the name and idea
% of effect comes from \docfm). It takes the same argument as
% |\DocInput|, the file's (path and) name with extension. (It
% \emph{has} |\DocInput| inside). It works properly if the input file
% doesn't contain explicit \<char1> (|^^A| is \acro{OK}).
%
% The effect of this command is typesetting of all the input file
% verbatim, with the code lines numbered and the \CSs automatically
% indexed (\pk{gmdoc.sty} options are in force).
%
%
% \subdivision{Package options}
%
% As many good packages, this also provides some options:
% \stanza
%
% Due to best \TeX\ documenting traditions the codelines will be
% numbered. But if the user doesn't wish that, \heshe\ may turn it off
% with the \TextUsage*{linesnotnum} option.
%
% However, if \heshe\ agrees to have the lines numbered, \heshe\ may wish
% to reset the counter of lines \himher self, e.g., when \heshe\
% documents many source files in one document. Then \heshe\ may wish
% the line numbers to be reset with every |{section}|'s turn for
% instance. This is the r\ocircum le of the \TextUsage*{uresetlinecount}
% option, which seems to be a~bit obsolete however, since the
% \TextCommonIndex\DocInclude|\DocInclude| command takes care of
% a~proper reset.
%
% Talking of line numbering further, a~tradition seems to exist to
% number only the codelines and not to number the lines of
% commentary. That's the default behaviour of \pk{gmdoc} but, if
% someone wants the comment lines to be numbered too, which may be
% convenient for reference purposes, \heshe\ is
% provided the \TextUsage*{countalllines} option. This option switches
% things to use the \cs{inputlineno} primitive for codeline numbers so
% you get the numbers of the source file instead of number only of the
% codelines. Note however, that there are no hypertargets made to the
% narration lines and the value of \cs{ref} is the number of the most
% recent codeline.
%
% Moreover, if \heshe\ wants to get the narration lines' number
% printed, there is the starred version of that option,
% \TextUsage*{countalllines*}. I~imagine someone may use it for
% debug. This option is not finished in details, it causes errors with
% \cs{addvspace} because it puts a~hyperlabel at every line. When it
% is in force, all the index entries are referenced
% with the line numbers and
%% ^^A(
% {\LineNumFont 441}\, the narration acquires a~bit biblical look ;-),
% {\LineNumFont 442}\,as shown in this short example. This option is intended
% {\LineNumFont 443}\,for the draft versions and it is not perfect (as if anything
% {\LineNumFont 444}\,in this package was). As you see, the lines
% {\LineNumFont 445}\,are typeset continuously with the numbers printed.
%
% \stanza
% By default the \pk{makeidx} package is loaded and initialized and
% the \CSs occurring in the code are automatically
% (hyper)indexed thanks to the \pk{hyperref} package. If the user
% doesn't wish to index anything, she should use the
% \TextUsage*{noindex} option.
%
% The index comes two possible ways: with the line numbers (if the
% lines are numbered) and that's the default, or with the page numbers, if
% the \TextUsage*{pageindex}\label{indexoptions} option is set.
%
% The references in the change history are of the same: when index is
% line number, then the changes history too.
%
% \stanza
% By default, \gmdoc\ excludes some 300 \CSs from being indexed.
% They are the most common \CSs, \LaTeX\ internal
% macros and \TeX\ primitives. To learn what \CSs are
% excluded actually, see lines \ref{DIE1}--\ref{DIE2}.
%
% If you don't want all those exclusions, you may turn them off with
% the \TextUsage*{indexallmacros} option.
%
% If you have ambiguous feelings about whether to let the default
% exclusions or forbid them, see p.\,\pageref{IEIDeclarations} to
% feed this ambiguity with a~couple of declarations.
%
% \stanza
% In \docfm\ package there's a~default behaviour of putting marked
% macro's or environment's name to a~marginpar. In the standard
% classes it's allright but not all the classes support marginpars.
% That is the reason why this package enables marginparing when in
% standard classes, enables or disables it due to the respective option
% when with Marcin Woli\nacute ski's classes and in any case provides the
% options \TextUsage*{withmarginpar} and \TextUsage*{nomarginpar}. So, in
% non-standard classes the default behaviour is to disable marginpars.
% If the marginpars are enabled in \pk{gmdoc}, it will put marked
% control sequences and environments into marginpars (see
% \gmiflink[textUsage]{\cs{TextUsage} etc.}). These options do
% not affect common using marginpars, which depends on the
% documentclass.
%
% \stanza
% My suggestion is to make the spaces in the code visible except the
% leading ones and that's the default. But if you wish all the code
% spaces to be blank, I~give the option \TextUsage*{codespacesblank}
% reluctantly. Moreover, if you wish the code spaces to be blank only
% in some areas, then there's \TextUsage\CodeSpacesBlank\ declaration
% (\acro{OCSR}).
%
% Another space formatting option is \TextUsage*{codespacesgrey}
% suggested by Will Robertson. It makes the spaces of code visible
% only not black but grey. The name of their colour is
% |visspacesgrey| and by default it's defined as |{gray}{.5}|, you
% can change it with \pk{xcolor}'s |\definecolor|. There is also an
% \acro{OCSR} declaration \TextUsage\CodeSpacesGrey.
%
% If for any reason you wish the code spaces blank in general and
% visible and grey in \env{verbatim*}s, use the declaration
% \TextUsage\VisSpacesGrey\ of the \pk{gmverb} package. If you like
% a~little tricks, you can also specify
% |codespacesgrey, codespacesblank| in \pk{gmdoc} options (in this
% order).
%
% \subdivision{The packages required}
%
% \pk{gmdoc} requires (loads if they're not loaded yet) some other
% packages of mine, namely \pk{gmutils}, \pk{gmverb}, analogous to
% Frank Mittelbach's \pk{shortvrb}, and \pk{gmiflink} for conditional
% making of hyperlinks. It also requires \pk{hyperref},
% \pk{multicol}, \pk{color} and \pk{makeidx}.
%
% The \pk{gmverb}\gmdmarginpar{\pk{gmverb}} package redefines the
% |\verb| command and the \env{verbatim} environment in such a~way that
% | |, |{| and |\| are breakable, the first with no `hyphen' and the
% other two with the comment char as a~hyphen, i.e.,
% |{|\<subsequent text>|}| breaks into
% |{%|\\ \<subsequent text>|}| ^^A] bal. braces for Emacs
% and \<text>|\mylittlemacro| breaks into \<text>|%|\\
% |\mylittlemacro|.
%
% As the standard \LaTeX\ one, my |\verb| issues an error when a~line
% end occurs in its scope. But, if you'd like to allow line ends in
% short verbatims, there's the \TextUsage\verbeolOK\ declaration. The
% plain |\verb| typesets spaces blank and |\verb*| makes them visible,
% as in the standard version(s).
%
% Moreover, \pk{gmverb} provides the \TextUsage\MakeShortVerb\
% declaration
% that takes a~one-char control sequence as the only argument and
% turns the char used into a~short verbatim delimiter, e.g., after
% \[\verb+\MakeShortVerb*\|+\]
% (as you see, the declaration has the starred version, which is for
% visible spaces, and non-starred for blank spaces) to get
% |\mylittlemacro| you may type \verb+|\mylittlemacro|+ instead of
% |\verb+\mylittlemacro+|. Because the char used in the last example
% is my favourite and is used this way by DEK in \TeXbook's
% format, \pk{gmverb} provides a~macro \TextUsage\dekclubs\ that
% expands to the example displayed above.
%
% Be careful because such active chars may interfere with other
% things, e.g., the \verb+|+ with the vertical line marker in \env{tabular}s and
% with the \pk{tikz} package. If this happens, you can declare e.g.,
% \TextUsage\DeleteShortVerb\verb+\|+ and the previous meaning of the
% char used shall be restored.
%
% One more difference between \pk{gmverb} and \pk{shortvrb} is that
% the chars |\active|ated by |\MakeShortVerb|, behave as if they were
% `other' in math mode, so you may type e.g., \verb+$k|n$+ to get
% $k|n$ etc.
% \stanza
%
% The \pk{gmutils}\gmdmarginpar{\pk{gmutils}} package provides a~couple of
% macros similar to some basic \LaTeXpar\ ones, rather strictly
% technical and (I~hope) tricky, such as |\afterfi|, |\ifnextcat|,
% |\addtomacro| etc. It's this package that provides the macros for
% formatting of names of macros and files, such as |\cs|, |\marg|, |\pk|
% etc. \stanza
%
% The \gmdoc\ package uses a~lot of hyperlinking possibilities
% provided by \pk{hyperref}\gmdmarginpar{\pk{hyperref}} which is therefore
% probably the most important package required. The recommended
% situation is that the user loads \pk{hyperref} package with \hisher\
% favourite options \emph{before} loading \pk{gmdoc}.
%
% If \heshe\ does not, \pk{gmdoc} shall load it with \emph{my}
% favourite options.
%
% To avoid an error if a~(hyper)referenced label does not exist,
% \pk{gmdoc} uses the \pk{gmiflink}\gmdmarginpar{\pk{gmiflink}}
% package. It works e.g., in the index when the codeline numbers have
% been changed: then they are still typeset, only not as hyperlinks
% but as a~common text.
%
% \stanza
% To typeset the index and the change history in balanced columns
% \gmdoc\ uses the \gmdmarginpar{\pk{multicol}}\pk{multicol} package
% that seems to be standard these days.
%
% Also the \gmdmarginpar{\pk{color}}\pk{multicol} package, required to
% define the default colour of the hyperlinks, seems to be standard
% already, and \pk{makeidx}.
%
% \subdivision[Automatic marking of definitions]{^^B
% \gmhypertarget{Automatic marking of definitions}}
% \gmdoc\ implements automatic detection of a~couple of
% definitions. By default it detects all occurrences of the following
% commands in the code:
%\begin{enumerate}^^B
%\item \label{def type}
% |\def|,
% |\newcount|,
% |\newdimen|,
% |\newskip|,
% |\newif|,
% |\newtoks|,
% |\newbox|,
% |\newread|,\\
% |\newwrite|,
% |\newlength|,
% |\newcommand(*)|,
% |\renewcommand(*)|,
% |\providecommand(*)|,
% |\DeclareRobustCommand(*)|,
% |\DeclareTextCommand(*)|,\\
% |\DeclareTextCommandDefault(*)|,
% |\DeclareDocumentCommand|,
%
% \item
% |\newenvironment(*)|,
% |\renewenvironment(*)|,
% |\DeclareOption(*)|,
%^^A~ |%\DeclareDefining*\@namedef|
%
% \item \label{newctr}
% |\newcounter|,
%
% \xdef\gmdenumi{\arabic{enumi}}
% \end{enumerate}
%
% of the \pk{xkeyval} package:
% \begin{enumerate}\setcounter{enumi}{\gmdenumi}^^B
% \item\label{dk type} |\define@key|,
% |\define@boolkey|,
% |\define@choicekey|,
% |\DeclareOptionX|,
%
% \xdef\gmdenumi{\arabic{enumi}}
% \end{enumerate}
%
% and of the \pk{kvoptions} package:
% \begin{enumerate}\setcounter{enumi}{\gmdenumi}^^B
% \item \label{DSOption}
% |\DeclareStringOption|,
% |\DeclareBoolOption|,
% |\DeclareComplementaryOption|,\\
% |\DeclareVoidOption|.
% \end{enumerate}\par
%
% What does `detects' mean? It means that the main argument of
% detected command will be marked as defined at this point,
% i.e.\ thrown to a~margin note and indexed with a~`definition' entry.
% Moreover, for the definitions \ref*{newctr}--\ref*{DSOption} an
% alternate index entries will be created: of the \CSs uderlying those
% definitions, e.g. |\newcounter{foo}| in the
% code will result in indexing |foo| and |\c@foo|.
%
% If you want to add detection of a~defining command not listed above,
% use the \TextUsage\DeclareDefining\ declaration. It comes in two
% flavours: `saut\eacute' and with star. The `saut\eacute' version
% (without star and without an optional argument) declares a~defining
% command of the kind of |\def| and |\newcommand|: its main argument,
% whether wrapped in braces or not, is a~\CS. The starred version (without
% the optional argument) declares a~defining command of the kind of
% |\newenvironment| and |\DeclareOption|: whose main mandatory
% argument is text. Both versions provide an optional argument in
% which you can set the keys.
%
% Probably the most important key is \TextUsage*{type}. Its default
% value is |cs| and that is set in the `saut\eacute' version. Another
% possible value is |text| and that is set in the starred version. You
% can also set three other types (any keyval setting of the type
% overrides the default and `starred' setting): |dk|, |dox| or |kvo|.
%
% |dk| stands for |\define@key| and is the type of \pk{xkeyval}
% definitions of keys (group \ref{dk type} commands). When detected,
% it scans furher code for an optional \arg[KVprefix], mandatory
% \arg{KVfamily} and mandatory \arg{key name}. The default \<KVprefix>
% is |KV|, as in \pk{xkeyval}.
%
% |dox| stands for |\DeclareOptionX| and launches scanning for an
% optional \arg[KVprefix], optional |<|\<KVfamily>|>| and mandatory
% \arg{option name}. Here the default \<KVprefix>
% is also |KV| and the default \<KVfamily> is the input file name. If
% you want to set another default family (e.g. if the code of
% \file{foo.sty} actually is in file \file{bar.dtx}), use
% \TextUsage\DeclareDOXHead\arg{KVfamily}. This declaration has an
% optional first argument that is the default \<KVprefix> for
% |\DeclareOptionX| definitions.
%
% |kvo| stands for the \pk{kvoptions} package by Heiko Oberdiek. This
% package provides a~handful of option defining commands (the group
% \ref{DSOption} commands). Detection of such a~command launches
% a~scan for mandatory \arg{option name} and alternate indexing of
% a~\CS |\|\<KVOfamily>|@|\<optionname>. The default \<KVOfamily> is
% the input file name. Again, if you want to set something else, you
% are given the \TextUsage\DeclareKVOFam\arg{KVOfamily} that sets the
% default family (and prefix: \<KVOfamily>|@|) for all the commands of
% group \ref{DSOption}.
%
% Next key recognized by |\DeclareDefining| is \TextUsage*{star}. It
% determines whether the starred version of a~defining command should
% be taken into account. For example, |\newcommand| should be
% declared with |[star=true]| while |\def| with
% |[star=false]|. You can also write just |[star]| instead of
% |[star=true]|. It's the default if the |star| key is omitted.
%
% There are also \TextUsage*{KVpref} and \TextUsage*{KVfam} keys if
% you want to redeclare the \pk{xkeyval} definitions with another
% default prefix and family.
%
% For example, if you wish |\@namedef| to be detected (the original
% \LaTeX\ version), declare
% \[|\DeclareDefining*[star=false]\@namedef|\] or
% \[|\DeclareDefining[type=text,star=false]\@namedef|\] (as stated
% above, |*| is equivalent |[type=text]|).
%
% On the other hand, if you want some of the commands listed above
% \emph{not} to be detected, write \TextUsage\HideDefining\<command>
% in the commentary. If both \<command> and \<command*> are detected,
% then both will be hidden. |\HideDefining| is always |\global|. Later
% you can resume detection of \<command> and \<command*> with
% \TextUsage\ResumeDefining\<command> which is always |\global|
% too. Moreover, if you wish to suspend automatic detection of the
% defining \<command> only once (the next occurrence), there is
% |\HideDefining*| which suspends detection of the next occurrence of
% \<command>. So, if you wish to `hide' |\providecommand*| once, write
% \[|\HideDefining*\providecommand*|\]
%
% If you wish to turn entire detection mechanism off, write
% \TextUsage\HideAllDefining\ in the narration layer. Then you can
% resume detection with \TextUsage\ResumeAllDefining. Both
% declarations are |\global|.
%
% The basic definition command, |\def|, seems to me a~bit
% ambiguous. Definitely \emph{not always} it defines important
% macros. But first of all, if you |\def| a~\CS excluded from indexing (see
% section \gmiflink{Index ex/inclusions}), it will not be marked even
% if detection of |\def| is on. But if the |\def|'s argument is not
% excluded from indexing and you still don't want it to be marked at
% this point, you can write |\HideDefining*\def| or \TextUsage\UnDef\
% for short.
%
% If you don't like |\def| to be detected more times, you may write
% |\HideDefining\def| of course, but there's a~shorthand for this:
% \TextUsage\HideDef\ which has the starred version
% \TextUsage\HideDef*\ equivalent |\UnDef|. To resume detection of
% |\def| you are provided also a~shorthand, \TextUsage\ResumeDef\ (but
% |\ResumeDefining\def| also works).
%
% If you define things not with easily detectable commands, you can
% mark them `manually', with the |\Define| declaration described in
% the next section.
%
%
%
% \subdivision[Manual marking of the macros and environments]{^^B
% \gmhypertarget{Manual Marking of the Macros and Environments}}
%
% The concept (taken from \docfm ) is to index virtually all the
% control sequences occurring in the code. \gmdoc\ does that by
% default and needs no special command. (See below about exluding some
% macros from being indexed.)
%
% The next concept (also taken from \docfm) is to ditinguish some
% occurrences of some control sequences by putting such a~sequence
% into a~marginpar and by special formatting of its index entry. That
% is what I~call marking the macros.
% \gmdoc\ provides also
% a~possibility of analogous marking for the environments' names and
% other sequences such as |^^A|.
%
% This package provides two kinds of special formatting of the
% index entries: `usage', with the reference number italic by default,
% and `def' (in \docfm\ called `main'), with the reference number
% roman (upright) and underlined by default. All the reference
% numbers, also those with no special formatting, are made hyperlinks
% to the page or the codeline according to the respective indexing
% option (see p.\,\pageref{indexoptions}).
%
% The macros and environments to be marked appear either in the code
% or in the commentary. But all the definitions appear in the code,
% I~suppose. Therefore the `def' marking macro is provided only for
% the code case. So we have\gmhypertarget[textUsage]{} the
% \TextUsage\Define, \TextUsage\CodeUsage\ and
% \TextUsage\TextUsage\ commands.
%
% All three take one argument and all three may be starred. The
% non-starred versions are intended to take a~control sequence as the
% argument and the starred to take whatever (an
% environment name or a~|^^A|-like and also a~\CS).
%
% You don't have to bother whether |@| is a~letter while documenting
% because even if not, these commands do make it a~letter, or more
% precisely, they execute \TextUsage\MakePrivateLetters\ whatever
% it does: At the default settings this command makes |*|
% a~letter, too, so a~starred version of a~command is a~proper argument
% to any of the three commands unstarred.
%
% The |\Define| and |\CodeUsage| commands, if unstarred, mark the next
% scanned occurrence of their argument in the code. (By `scanned
% occurrence' I~mean a~situation of the \CS having been scanned in the
% code which happens iff its name was preceded by the char declared as
% |\CodeEscapeChar|). The starred versions of those commands mark just
% the next codeline and don't make \TeX\ looks for the scanned
% occurrence of their argument (which would never happen if the
% argument is not a~\CS). Therefore, if you want to mark a~definition
% of an environment \env{foo}, you should put
% \[|%\Define*{foo}|\]
% right before the code line
% \[|\newenvironment{foo}{%|\]^^A]
% i.e., not separated by another code line. The starred versions of
% the |\Code...| commands are also intended to mark implicit
% definitions of macros, e.g., |\Define*\@foofalse| before
% the line
% \[|\newif\if@foo|.\]
%
% They both are |\outer| to dicourage their use inside macros because
% they actually re|\catcode| before taking their arguments.
%
% The |\TextUsage| (one-argument) command is intended to mark usage of
% a~verbatim occurrence of a~\TeX\ object in the commentary. Unlike
% |\CodeUsage| or |\Define|, it typesets its argument which means among others\ that
% the marginpar appears usually at the same line as the text you
% wanted to mark. This command also has the starred version primarily
% intended for the environments names, and secondarily for
% |^^A|-likes and \CSs, too. Currently, the most important difference
% is that the unstarred version executes |\MakePrivateLetters| while
% the starred does both |\MakePrivateLetters| and |\MakePrivateOthers|
% before reading the argument.
%
% If you consider the marginpars a~sort of sub(sub\dots)section
% marks, then you may wish to have a~command that makes a~marginpar of
% the desired \CS (or whatever) at the beginning of its description,
% which may be fairly far from the first occurrence of its
% object. Then you have the \TextUsage\Describe\ command which
% puts its argument in a~marginpar and indexes it as a~`usage' entry
% but doesn't print it in the text. It's |\outer|.
%
% All four commands just described put their (|\string|ed) argument
% into a~marginpar (if the marginpars are enabled) and create an index
% entry (if indexing is enabled).
%
% But what if you want just to make a~marginpar with macro's or
% environment's name? Then you have \TextUsage\CodeMarginize\
% to declare what to put into a~marginpar in the \TeX\ code (it's
% |\outer|) and
% \TextUsage\TextMarginize\ to do so in the commentary. According to
% the spirit of this part of the interface, these commands also take one
% argument and have their starred versions for strings other than
% control sequences.
%
% The marginpars (if enabled) are `reverse' i.e., at the left margin, and
% their contents is flush right and typeset in a~font declared with
% \TextUsage\marginpartt. By default, this declaration is |\let| to
% |\tt| but it may be advisable to choose a~condensed font if there is
% any. Such a~choice is made by \pk{gmdocc.cls} if the Latin Modern
% fonts are available: in this case \pk{gmdocc.cls} uses Latin Modern
% Typewriter Light Condensed.
%
% If you need to put something in a~marginpar without making it typewriter
% font, there's the \TextUsage\gmdmarginpar\ macro (that takes one
% and mandatory argument) that only flushes its contents right.
%
% \stanza
% On the other hand, if you don't want to put a~\CS (or another
% verbatim text) in a~marginpar but only to index it, then there are
% \TextUsage\DefIndex\ and \TextUsage\CodeUsgIndex\ to declare
% special formatting of an entry. The unstarred versions of these
% commands look for their argument's scanned occurrence in the code
% (the argument should be a~\CS), and the starred ones just take the
% next code line as the reference point. Both these commands are
% |\outer|.
%
% In the code all the control sequences (except the excluded ones, see
% below) are indexed by default so no explicit command is needed for
% that. But the environments and other special sequences are not and
% the two commands described above in their |*|ed versions contain the
% command for indexing their argument. But what if you wish to index
% a~not scanned stuff as a~usual entry? The \TextUsage\CodeCommonIndex*\
% comes in rescue, starred for the symmetry with the two previous commands
% (without |*| it just gobbles
% it's argument---it's indexed automatically anyway). It's |\outer|.
%
% Similarly, to index a~\TeX\ object occurring verbatim in the narrative, you
% have \TextUsage\TextUsgIndex\ and \TextUsage\TextCommonIndex\
% commands with
% their starless versions for a~\CS argument and the starred for all
% kinds of the argument.
%
% \stanza
% Moreover, as in \docfm, the \TextUsage*{macro} and
% \TextUsage*{environment} environments are provided. Both take one
% argument that should be a~\CS for \env{macro} and `whatever' for
% \env{environment}. Both add the |\MacroTopsep| glue before and after
% their contents, and put their argument in a~marginpar at the first
% line of their contents (since it's done with |\strut|, you should
% not put any blank line (|%|ed or not) between
% |\begin{macro/environment}| and the first line of the
% contents). Then \env{macro} commands the first scanned occurrence of
% its argument to be indexed as `def' entry and \env{environment}
% commands \TeX\ to index the argument as if it occurred in the next
% code line (also as `def' entry).
%
% Since it's possible that you define a~\CS implicitly i.e., in such a~way that it
% cannot be scanned in the definition (with |\csname...\endcsname|
% e.g.) and wrapping such a~definition (and description) in an
% \env{environment} environment would look misguidedly ugly, there's
% the \env{macro*} environment which \TeX nically is just an alias for
% \env{environment}.
%
% (To be honest, if you give a~\env{macro} environment a~non-\CS
% argument, it will accept it and then it'll work as
% \env{evironment}.)
%
%
% \subdivision[Index ex/inclusions]{\gmhypertarget{Index ex/inclusions}}
%
% It's\label{IEIDeclarations} understandable\footnote{After reading ^^B(
% \docfm's documentation ;-)\,.} that you don't
% want some control sequences to be indexed in your documentation. The
% \docfm\ package gives a~brilliant solution: the
% \TextUsage\DoNotIndex\ declaration. So do I
% (although here, \TeX nically it's done another
% way). It \acro{OCSR}. This declaration takes one
% argument consisting of a~list of control sequences not to be
% indexed. The items of this list may be separated with commas, as in
% \docfm , but it's not obligatory. The whole list should come in
% curly braces (except when it's one-element), e.g.,
% \[|\DoNotIndex{\some@macros,\are* \too\auxiliary\?}|\]
% (The spaces after the control sequences are ignored.)
% You may use as many |\DoNotIndex|es as you wish (about half as many as
% many \CSs may be declared, because for each \CS excluded from indexing
% a~special \CS is declared that stores the ban sentence).
% Excluding the same \CS more than once makes no problem.
%
% I~assume you wish most of \LaTeX\ macros, \TeX\ primitives etc.\ to
% be excluded from your index (as I~do). Therefore \pk{gmdoc} excludes
% some 300 \CSs by default. If you don't like it, just set the
% |indexallmacros| package option.
%
% On the third hand, if you like the default exclusions in general but
% wish to undo just a~couple of them, you are given \TextUsage\DoIndex\
% declaration (\acro{OCSR}) that removes a~ban on all the \CSs given in the
% argument, e.g.,
% \[|\DoIndex{\par \@@par \endgraf}|\]
%
% Moreover, you are provided the \TextUsage\DefaultIndexExclusions\
% and \Describe\UndoDefaultIndexExclusions^^A
% \cs{Un\-do\-Def\-ault\-Ind\-ex\-Ex\-cl\-us\-ions} declarations that
% act according to their names. You may use them in any configuration
% with the |indexallmacros| option. Both of these declarations \acro{OCSR}.
%
%
% \subdivision{The \ds\ directives}
%
% \gmdoc\ typesets the \ds\ directives and it does it quite likely as
% \docfm, i.e., with math sans serif font. It does it automatically
% whether you use the traditional settings or the new.
%
% Advised by my \TeX\ Guru, I~didn't implement the module nesting
% recognition (MW told it's not that important.)
%
% So far \gmhypertarget{verbatim mode directive} is only
% half-handled. That is, a~line beginning with |%<||<|\<END-TAG> will be
% typeset as a~\ds\ directive, but ^^A~The magical directive divided
% ^^A due to the doctex mode who saw it and coloured all after with
% ^^A~salmon.
% the closing line |%|\<END-TAG> will be not. It doesn't seem to be
% hard to implement, if I~only receive some message it's really useful
% for someone.
%
%^^A If you use the \TextUsage*{macrocode} environment, you don't have to
%^^A add any \CS to get \ds\ directive typeset properly: within
%^^A \env{macrocode} a~|<| sign in a~line commented out is read as
%^^A a~beginning of a~\ds\ directive, which means that \TeX\ looks
%^^A for the closing |>| or the very next |<| as the beginning of the
%^^A special directive |<<|\<any text till the end of line>. In the
%^^A standard case of |<|\<directive>|>| \TeX\ typesets any further text
%^^A in its line as the code.
%^^A
%^^A If you want the \ds\ directives to be typeset without
%^^A \env{macrocode}, you shoud declare \TextUsage\docstrips\ to make
%^^A \TeX\ read the |<| in the comment layer as the beginning of
%^^A a~\ds\ directive. In |\docstrips|'s scope the |<| signs
%^^A occurring in math mode remain just inequality signs.
%^^A
%^^A If you wish only the first |<| met to start a~\ds\ directive
%^^A typesetting, then there's \TextUsage\docstrip\ declaration that
%^^A restores the previous meaning of |<| after the first using it as
%^^A a~\ds\ directive opener.
%^^A
%^^A Both |\docstrips| and |\docstrip| \acro{OCSR} and, of course you may use
%^^A them as environments (which doesn't make much sense with the
%^^A latter, though).
%^^A
%
% \subdivision{The changes history}
%
% The \docfm's documentation reads:
% \begin{quotation}
% To maintain a change history within the file, the |\changes| command may
% be placed amongst the description part of the changed code. It takes three
% arguments, thus:
% \[|\changes{|\<version>|}{|\<\acro{YYYY}/\acro{MM}/\acro{DD} date>|}{|\<text>|}|\]
%
% The changes may be used to produce an auxiliary file (\LaTeX's
% |\glossary| mechanism is used for this) which may be printed after
% suitable formatting. The |\changes| [command] encloses the \<date> in
% parentheses and appends the \<text> to form the printed entry in
% such a change history [\dots\ obsolete remark ommitted].
%
% To cause the change information to be written out, include
% \TextUsage\RecordChanges\ in the driver['s preamble or just in the
% source file (\pk{gmdocc.cls} does it for you)]. To read in and print
% the sorted change history (in two
% columns), just put the \TextUsage\PrintChanges\ command as the last
% (commented-out, and thus executed during the documentation pass
% through the file) command in your package file [or in the driver].
% Alternatively, this command may form one of the arguments of the
% |\StopEventually| command, although a change history is probably not
% required if only the description is being printed. The command
% assumes that MakeIndex or some other program has processed the
% \pk{.glo} file to generate a~sorted \pk{.gls} file. You need
% a~special MakeIndex style file; a suitable one is supplied with
% \docfm\ [and \gmdoc], called \iffalse\pk{gglo.ist}\fi^^A
% [\dots\ \textbf{\pk{gmglo.ist}} for \gmdoc]. The \TextUsage\GlossaryMin,
% \TextUsage\GlossaryPrologue\ and \TextUsage\GlossaryParms\ macros are
% analagous to the |\Index...| versions [see
% sec.~\gmiflink[IndexParams]{The parameters}
% p.\,\pageref*{IndexParams}]. (The \LaTeX\ `glossary'
% mechanism is used for the change entries.)
% \end{quotation}
%
% In \gmdoc\ (unless you turn definitions detection off), you can put
% |\changes| after the line of definition of a~command to set the
% default argument of |\changes| to that command. For example,
%\begin{verbatim}
% \newcommand*\dodecaphonic{...}
% % \changes{v0.99e}{2007/04/29}{renamed from \cs{DodecaPhonic}}
%\end{verbatim}
% results with a~history (sub)entry:
% \deksmallskip
% v0.99e\par\hspace*{\parindent} (\dots)\par \hspace*{\parindent}
% \cs{dodecaphonic}: \par\hspace*{2\parindent} renamed from
% \cs{DodecaPhonic}, \textit{\arabic{page}}
% \deksmallskip
%
% Such a~setting is in force till the next definition and
% \emph{every} detected definition resets it.
%
% In \gmdoc\ |\changes| is |\outer|.
%
% \stanza
% As mentioned in the introduction, the glossary, the changes history
% that is, uses a~special MakeIndex style, \pk{gmglo.ist}. This style
% declares another set of the control chars but you don't have to
% worry: |\changes| takes care of setting them properly. To be
% precise, |\changes| executes \TextUsage\MakeGlossaryControls\ that
% is defined as
% \begin{verbatim}
% \def\actualchar{=} \def\quotechar{!}%
% \def\levelchar{>} \edef\encapchar{\xiiclub}
%\end{verbatim}
%
% Only if you
% want to add a~control character yourself in a~changes entry, to quote
% some char, that is (using level or encapsulation chars is not
% recommended since |\changes| uses them itself), use rather
% |\quotechar|.
%
% \stanza
% Before writing an entry to the \pk{.glo} file, |\changes|
% checks if the date (the second mandatory\equals the third argument)
% is later than the date stored in the counter
% \TextUsage*{ChangesStartDate}. You may set this counter with a
% \Describe\ChangesStart^^B
% \[|\ChangesStart{|\<version>|}{|\<year>|/|\<month>|/|\<day>|}|\]
% declaration.
%
% If the \env{ChangesStartDate} is set to a~date contemporary to \TeX\
% i.e., not earlier than September 1982\footnote{DEK in \textit{\TeX\ ^^B
% The Program} mentions that month as of \TeX\ Version 0 release.},
% then a~note shall appear at the beginning of the changes history
% that informs the reader of ommitting the earlier changes entries.
%
% If the date stored in \env{ChangesStartDate} is earlier than \TeX,
% no notification of ommitting shall be printed. This is intended for
% a~rather tricky usage of the changes start date feature: you may
% establish two threads of the changes history: the one for the users,
% dated with four digit year, and the other for yourself only, dated
% with two or three digit year. If you declare
% \[|\ChangesStart{|\<version?>|}{1000/00/00}|\]
% or so, the changes entries dated with less-than-four digit year
% shall be ommitted and no notification shall be issued of that.
%
% \stanza
%
% While scanning the \CSs in the code, \gmdoc\ counts them and prints
% the information about their number on the terminal and in
% \pk{.log}. Moreover, you may declare
% \TextUsage\CheckSum|{|\<number>|}| before the code and \TeX\ will
% inform you whether the number stated by you is
% correct or not, and what it is. As you guess, it's not my original idea
% but I~took it from \docfm.
%
% There it is provided as a~tool for
% testing whether the file is corrupted.
% My \TeX\ Guru says it's a~bit old-fashioned nowadays but I~like the
% idea and use it to document the file's growth. For this purpose
% \gmdoc\ types out lines like
% \begin{verbatim}
%% \chschange{v0.98j}{2006/10/19}{4372}
%% \chschange{v0.98j}{06/10/19}{4372}
%\end{verbatim}
% and you may place them at the beginning of the source file. Such
% a~line results in setting the check sum to the number contained in
% the last pair of braces and in making a~`general' changes entry that
% states the check sum for version \<first brace> dated \<second ^^B
% brace> was \<third brace>.
%
% \stanza
% There is also \TextUsage\toCTAN\arg{version}\arg{date}, a~shorthand
% for
% \[|\changes|\arg{version}\arg{date}|{put to \acro{CTAN} on |\<date>|}|\]
%
%
% \subdivision{The parameters}
%
% The \pk{gmdoc} package provides some parameters specific to
% typesetting the \TeX\ code:
%
% \stanza
% \TextUsage\stanzaskip\ is a~vertical space inserted when a~blank
% (code) line is met. It's equal |0.75\medskipamount| by default
% (with the \emph{entire} |\medskipamount|'s stretch- and
% shrinkability). Subsequent blank code lines do not increase this
% space.
%
% At the points where narration begins a~new line after the code or an
% inline comment and where a~new code line begins after the narration
% (that is not an inline comment), a~\TextUsage\CodeTopsep\ glue is
% added. At the beginning and the end of a~\env{macro} or
% \env{environment} environment a~|\MacroTopsep| glue is added. By
% default, these two skips are set equal |\stanzaskip|.
%
% The |\stanzaskip|'s value is assigned also to the display skips and
% to |\topsep|. This is done with the \TextUsage\UniformSkips\
% declaration executed by default. If you want to change some of those
% values, you should declare \TextUsage\NonUniformSkips\ in the
% preamble to discard the default declaration. (To be more precise, by
% default |\UniformSkips| is executed twice: when loading \gmdoc\ and
% again |\AtBeginDocument| to allow you to change |\stanzaskip| and
% have the other glues set due to it. |\NonUniformSkips| relaxes the
% |\UniformSkips|'s occurrence at |\begin{document}|.)
%
% If you want to add a~vertical space of |\CodeTopsep| (equal by
% default |\stanzaskip|), you are provided the \TextUsage\stanza\
% command. Similarly, if you want to add a~vertical space of the
% |\MacroTopsep| amount (by default also equal |\stanzaskip|), you are
% given the \TextUsage\chunkskip\ command. They both act analogously
% to |\addvspace| i.e., don't add two consecutive glues but put the
% bigger of them.
%
% Since |\CodeTopsep| glue is inserted automatically at each
% transition from the code (or code with an inline comment) to the
% narration and reverse, it may happen that you want not to add such
% a~glue exceptionally. Then there's the \TextUsage\nostanza\
% command. You can use it before narration to remove the vskip before
% it or after narration to suppress the vskip after it.
%
% \stanza
% The \TeX\ code is indented with the \TextUsage\CodeIndent\ glue
% and a~leading space increases indentation of the line by its
% (space's) width. The default value of |\CodeIndent| is 1.5\,em.
%
% There's also a~parameter for the indent of the narration,
% \TextUsage\TextIndent, but you should use it only in emergency
% (otherwise what would be the margins for?). It's 0\,sp by default.
%
% \stanza
% By default, the end of a~|\DocInput| file is marked with
%
%`\EOFMark'
%
% \noindent given by the
% \TextUsage\EOFMark\ macro.
%
% \stanza
% If you do use the \eTeX's primitive \TextUsage\everyeof, be sure
% the contents of it begins with |\relax| because it's the token that
% stops the main macro scanning the code.
% \stanza
%
% The crucial concept of \gmdoc\ is to use the line end character
% as a~verbatim group opener and the comment char, usually the |%|, as
% its delimiter. Therefore the `knowledge' what char starts
% a~commentary is for this package crucial and utterly
% important. The default assumption is that you use |%| as we
% all do. So, if you use another character, then you should declare it
% with \TextUsage\CodeDelim\ typing the desired char
% preceded by a~backslash, e.g., |\CodeDelim\&|\,. (As just
% mentioned implicitly, \possfil|\CodeDelim\%| is declared by
% deafult.)
%
% This declaration is always global so when- and wherever you change
% your mind you should express it with a~new |\CodeDelim|
% declaration.
%
% The starred version of |\CodeDelim| changes also the verb
% `hyphen', the char appearing at the verbatim line breaks that is.
%
% \stanza
% Talking of special chars, the escape char, |\| by default, is also
% very important for this package as it marks control sequences and
% allows automatic indexing them for instance. Therefore, if you for
% any reason choose another than |\| character to be the escape char,
% you should tell \pk{gmdoc} about it with the
% \TextUsage\CodeEscapeChar\ declaration. As the previous one, this too
% takes its argument preceded by a~backslash,
% e.g., |\CodeEscapeChar\!|. (As you may deduct from the above,
% |\CodeEscapeChar\\| is declared by default.)
%
% The tradition is that in the packages |@| char is a~letter
% i.e., of catcode \catletter. Frank Mittelbach in \docfm\ takes into account
% a~possibility that a~user wishes some other chars to be letters,
% too, and therefore he (F.M.) provides the
% \TextUsage\MakePrivateLetters\ macro.
% So do I~and like in \docfm, this macro makes |@| sign a~letter. It
% also makes |*| a~letter in order to cover the starred versions
% of commands.
%
% Analogously but for a~slightly different purpose, the
% \TextUsage\AddtoPrivateOthers\ macro is provided here. It adds its
% argument, which is supposed to be a~one-char \CS, to the
% |\doprivateothers| list, whose r\ocircum le is to allow some special chars
% to appear in the marking commands' arguments (the commands described
% in section \gmiflink{Macros for marking the macros}). The default
% contents of this list is | | (the space) and |^| so you may mark the
% environments names and special sequences like |^^A| safely. This
% list is also extended with every char that is |\MakeShortVerb|ed.
% (I~don't see a~need of removing chars from this list, but if
% you do, please let me know.)
%
% \stanza
% The line numbers (if enabled) are typeset in the
% \TextUsage\LineNumFont\ declaration's scope, which
% is defined as |{\normalfont\tiny}| by default. Let us also remember, that for
% each counter there is a~|\the|\<counter> macro available. The
% counter for the line numbers is called \TextUsage*{codelinenum} so
% the macro printing it is |\thecodelinenum|. By default we don't
% change its \LaTeX's definition which is equivalent
% |\arabic{codelinenum}|.
%
%
% Three more parameter macros, are \TextUsage\IndexPrefix,
% \TextUsage\EntryPrefix\ and \TextUsage\HLPrefix. All three are
% provided with the account of including multiple files in one
% document.
% They are equal (almost) |\@empty| by default. The first may store
% main level index entry of which all indexed macros and environments
% would be subentries, e.g., the name of the package. The third may or
% even should store a~text to distinguish equal codeline numbers of
% distinct source files. It may be the file name too, of course. The
% second macro is intended for another concept, namely the one from
% \pk{ltxdoc} class, to distinguish the codeline numbers from
% different files \emph{in the index} by the file marker. Anyway, if
% you document just one file per document, there's no need of
% redefining those macros, nor when you input multiple files with
% |\DocInclude|.
%
% \pk{gmdoc} automatically indexes the control sequences
% occurring in the code. Their index entries may be `common' or
% distinguished in two (more) ways. The concept is to distinguish the
% entries indicating the \emph{usage} of the \CS and the entries
% indicating the \emph{definition} of the \CS.
%
% The special formattings of `usage' and `def' index entries are
% determined by \Describe\UsgEntry\cs{Usg\-Ent\-ry} and \TextUsage\DefEntry\
% one-parameter macros (the parameter shall be substituted with the
% reference number) and by default are defined as |\textit| and
% |\underline| respectively (as in \docfm).
%
% There's one more parameter macro, \TextUsage\CommonEntryCmd\ that
% stores the name of the encapsulation for the `common' index entries
% (not special) i.e., a~word that'll become a~\CS that will be put
% before an entry in the \file{.ind} file. By default it's defined as
% |{relax}| and a~nontrivial use of it you may see in the source of
% chapter \ref*{Driver}, where
% |\def\CommonEntryCmd{UsgEntry}| makes all the index entries of the
% driver formatted as `usage'.
%
% \stanza
% ^^A~Index Parameters
% \gmhypertarget[IndexParams]{The index} comes in a~\env{multicols}
% environment whose columns number is determined by the
% \TextUsage*{IndexColumns} counter set by default to~3. To save space,
% the index begins at the same page as the previous text provided
% there is at least \TextUsage\IndexMin\ of the page height free. By
% default, |\IndexMin|\equals\the\IndexMin.
%
% The text put at the beginning of the index is declared with
% a~one-argument
% \Describe\IndexPrologue\cs{Ind\-ex\-Pro\-lo\-gue}. Its default text
% at current index option you may \gmiflink[DIPrologue]{admire} on
% page \pageref*{DIPrologue}.^^A
% \AtDIPrologue{\raisebox{4cm}[0sp][0sp]{\gmhypertarget[DIPrologue]}}
% Of course, you may write your own |\IndexPrologue{|\<brand new ^^B
% index prologue>|}|, but if you like the default and want only to add
% something to it, you are provided \TextUsage\AtDIPrologue\
% one-argument declaration that adds the stuff after the default
% text. For instance, I~used it to add a~label and hypertarget that is
% referred to two sentences earlier.
%
% By default the colour of the index entry hyperlinks is set black to
% let Adobe Reader work faster. If you don't want this,
% \Describe\IndexLinksBlack|\let\IndexLinksBlack\relax|. That leaves
% the index links colour alone and hides
% the text about black links from the default index prologue.
%
% Other index parameters are set with the \TextUsage\IndexParms\ macro
% defined in line \gmifref[IndexParms]{sixty-odd} of the code. If you
% want to change some of them, you don't have to use
% |\renewcommand*\IndexParms| and set all of the parameters: you may
% \Describe\gaddtomacro^^A
% |\gaddtomacro\IndexParms{|\<only the desired changes>|}|.
% (|\gaddtomacro| is an alias for \LaTeX's |\g@addto@macro| provided
% by \pk{gmutils}.)
%
% At the default \gmdoc\ settings the \file{.idx} file is prepared for
% the default settings of MakeIndex (no special style). Therefore the
% index control chars are as usual. But if you need to use other
% chars as MakeIndex controls, know that they are stored in the four
% macros: \TextUsage\actualchar, \TextUsage\quotechar,
% \TextUsage\levelchar\ and \TextUsage\encapchar\ whose meaning you
% infer from their names. Any redefinition of them \emph{should be ^^B
% done in the preamble} because the first usage of them takes place at
% |\begin{document}| and on it depends further tests telling \TeX\
% what characters of a~scanned \CS name it should quote before
% writing it to the \pk{.idx} file.
%
% \stanza
% Frank Mittelbach in \docfm\ provides the |\verbatimchar| macro to
% (re)define the |\verb|'s delimiter for the index entries of the
% scanned \CS names etc. \gmdoc\ also uses \TextUsage\verbatimchar\ but
% defines it as |{&}|. Moreover, a~macro that wraps a~\CS name in |\verb|
% checks whether the wrapped \CS isn't |\&| and if it is, |$| is taken
% as the delimiter. So there's hardly chance that you'll need to
% redefine |\verbatimchar|.
%
% So strange delimiters are chosen deliberately to allow any `other'
% chars in the environments names.
%
% \stanza
% There's a~quadratus of commands taken from \docfm:
% \TextUsage\StopEventually, \TextUsage\Finale,\\
% \TextUsage\AlsoImplementation\ and \TextUsage\OnlyDescription\ that
% should be
% explained simultaneously (in a~polyphonic song
% e.g.).
%
% The |\OnlyDescription| and |\AlsoImplementation|
% declarations are intended to exclude or include the code part from
% the documentation. The point between the description and the
% implementation part should be marked with |\StopEventually{|\<the ^^B
% stuff to be executed anyway>|}| and |\Finale| should be typed
% at the end of file. Then |\OnlyDescription| defines
% |\StopEventually| to expand to its argument followed by |\endinput|
% and\\ |\AlsoImplementation| defines |\StopEventually| to do nothing
% but pass its argument to |\Finale|.
%
%
% \subdivision{The narration macros}
%
% To print the control sequences' names you have the \TextUsage\verb\ macro and
% its `shortverb' version whatever you define (see the \pk{gmverb}
% package).
%
% For short verbatim texts in the inline comments \pk{gmdoc} provides
% the \TextUsage\inverb\<charX>\dots\<charX> (the name stands for
% `inline verbatim') command that redefines the \pk{gmverb} breakables
% to break with |%| at the beginning of the
% lower line to avoid mistaking such a~broken verbatim commentary text
% for the code.
%
% But nor |\verb(*)| neither |\inverb| will work if
% you put them in an argument of another macro. For such a~situation,
% or if you just prefer, \gmdoc\ (\pk{gmutils}) provides a~robust command
% \TextUsage\cs, which takes one obligatory argument, the macro's name
% without the backslash, e.g., |\cs{mymacro}| produces
% \cs{mymacro}. I~take account of a~need of printing some other text
% verbatim, too, and therefore |\cs| has the first argument
% optional, which is the text to be typeset before the mandatory
% argument. It's the backslash by default, but if you wish to typeset
% something without the |\|, you may write {\verbeolOK |\cs[]{not
% a~macro}|}. Moreover, for typesetting the environments' names,
% \gmdoc\ (\pk{gmutils})
% provides the \TextUsage\env\ macro, that prints its argument
% verbatim and without a~backslash, e.g., |\env{an environment}|
% produces \env{an environment}.
%
% For usage in the inline comments there are \TextUsage\incs\ and
% \TextUsage\inenv\ commands that take analogous arguments and precede
% the typeset command and environment names with a~|%| if at the
% beginning of a~new line.
%
% And for line breaking at |\cs| and |\env| there is
% \TextUsage\nlpercent\ to ensure |%| if the line breaks at the
% beginning of a~|\cs| or |\env| and \TextUsage\+ to use inside their
% argument for a~discretionary hyphen that'll break to - at the end of
% the upper line and |%| at the beginning of the lower line. By
% default hyphenation of |\cs| and |\env| arguments is off, you can
% allow it only at |\-| or |\+|.
%
% \stanza By default the multiline inline comments are typeset with
% a~hanging indent (that is constant relatively to the current
% indent of the code) and justified. Since vertical alignment is
% determined by the parameters as they are at the moment of \cs{par},
% no one can set the code line to be typeset ragged right (to break
% nicely if it's long) and the following inline comment to be
% justified. Moreover, because of the hanging indent the lines of
% multiline inline comments are relatively short, you may get lots of
% overfulls. Therefore there is a~Boolean switch \TextUsage*{ilrr}
% (\acro{OCSR}),
% whose name stands for `InLine RaggedRight' and the inline comments
% (and their codelines) are typeset justified in the scope of
% |\ilrrfalse| which is the default.
% When you write |\ilrrtrue|, then all inline comments in its scope
% (and their codelines) will be typeset ragged right
% (and still with the hanging indent). Moreover, you are provided
% \TextUsage\ilrr\ and \TextUsage\ilju\ commands that set |\ilrrtrue|
% and |\ilrrfalse| for the current inline comment only. Note you can
% use them anywhere within such a~comment, as they set |\rightskip|
% basically. |\ilrr| and |\ilju| are no-ops in the standalone narration.
%
%
%
% \stanza
% To print packages' names sans serif there is a~\TextUsage\pk\
% one-argument command, and the \TextUsage\file\ command intended for
% the filenames.
%
% Because we play a~lot with the |\catcode|s here and want to talk
% about it, there are \TextUsage\catletter, \TextUsage\catother\ and
% \TextUsage\catactive\ macros that print \catletter, \catother\ and
% \catactive\ respectively to concisely mark the most used char
% categories.
%
% I~wish my self-documenting code to be able to be typeset each
% package separately or several in one document. Therefore I~need some
% `flexible' sectioning commands and here they are:
% \TextUsage\division, \TextUsage\subdivision\ and
% \TextUsage\subsubdivision\ so far, that by default are |\let| to be
% |\section|, |\subsection| and |\subsubsection| respectively.
%
% \stanza
% One more kind of flexibility is to allow using \pk{mwcls} or the
% standard classes for the same file. There was a~trouble with the
% number and order of the optional arguments of the original
% \pk{mwcls}'s sectioning commands.
%
% It's resolved in \pk{gmutils} so you are free at this point, and
% even more free than in the standard classes: if you give
% a~sectioning command just one optional argument, it will be the
% title to toc and to the running head (that's standard in
% scls\footnote{See \pk{gmutils} for some subtle details.}). If you give
% \emph{two} optionals, the first will go to the running head and the
% other to toc. (In both cases the mandatory argument goes only to the
% page).
%
% \stanza
% If you wish the |\DocInclude|d files make other sectionings than the
% default, you may declare \TextUsage\SetFileDiv|{|\<sec name without ^^B
% backslash>|}|.
%
% \stanza
%
% \pk{gmdoc.sty} provides also an environment \TextUsage*{gmlonely}
% to wrap some text you think you may want to skip some day. When
% that day comes, you write \TextUsage\skipgmlonely\ before the
% instances of \env{gmlonely} you want to skip. This declaration has
% an optional argument which is for a~text that'll appear in(stead of) the
% first \env{gmlonely}'s instance in every |\DocInput| or
% |\DocInclude|d file within |\skipgmlonely|'s scope.
%
% An example of use you may see in this documentation: the repeated
% passages about the installation and compiling the documentation are
% skipped in further chapters thanks to it.
%
% \stanza
% \gmdoc\ (\pk{gmutils}, to be precise) provides some \TeX-related
% logos:\\
% \Describe\AmSTeX typesets \AmSTeX,\\
% \Describe\BibTeX\BibTeX,\\
% \Describe\SliTeX\SliTeX,\\
% \Describe\PlainTeX\PlainTeX,\\
% \Describe\Web\Web,\\
% \Describe\TeXbook\TeXbook,\par^^A~without par there raised error
% ^^A `too many unprocessed floats'.
% \noindent\Describe\TB\TB\par
% \noindent\Describe\eTeX\eTeX,\\
% \Describe\pdfeTeX\pdfeTeX\\
% \Describe\pdfTeX\pdfTeX\\
% \Describe\XeTeX\XeTeX\ (the first |E| will be reversed if the
% \pk{graphics} package is loaded or \XeTeX\ is at work) and\\
% \Describe\LaTeXpar\LaTeXpar.\par
%
% \noindent\Describe\ds\ds\ not quite a~logo, but still convenient.
%
% \stanza
% The \TextUsage*{copyrnote} environment is provided to format the
% copyright note flush left in |\obeylines|' scope.
%
% \stanza
%
% To put an arbitrary text into a~marginpar and have it flushed right
% just like the macros' names, you are provided the
% \TextUsage\gmdmarginpar\ macro that takes one mandatory argument
% which is the contents of the marginpar.
%
% \stanza
% To make a~vertical space to separate some piece of text you are
% given two macros: \TextUsage\stanza\ and \TextUsage\chunkskip. The
% first adds |\stanzaskip| while the latter |\MacroTopsep|. Both of
% them take care of not cumulating the vspaces.
%
%
%\stanza
% The \TextUsage*{quotation} environment is redefined just to enclose
% its contents in double quotes.
%
% If you don't like it, just call
% |\RestoreEnvironment{quotation}| after loading \pk{gmdoc}.
% Note however that other environments using
% \env{quotation}, such as \env{abstract}, keep their shape.
%
% \stanza
%
% The \TextUsage\GetFileInfo|{|\<file name with extension>|}| command
% defines \TextUsage\filedate, \Describe\fileversion\cs{fil\-e\-vers\-ion} and
% \TextUsage\fileinfo\ as the respective pieces of the info (the
% optional argument) provided
% by \cs{Pro\-vid\-es\-Class/Pack\-age/Fi\-le} declarations. The
% information of the file you process with \gmdoc\
% is provided (and therefore getable) if the file is also loaded (or
% the |\Provide...| line occurs in a~|\StraightEOL| scope).
%
% If the input file doesn't contain |\Provides...| in the code layer,
% there are commands \TextUsage\ProvideFileInfo|{|\<file name with ^^B
% extension>|}[|\<info>|]|. (\<info> should consist of:
% \<year>|/|\<month>|/|\<day>| |\<version number>| |\<a~short note>.)
%
% Since we may documentally input files that we don't load, \docfm\ in
% \pk{gmdoc} e.g., we provide a~declaration to be put (in the comment
% layer) before the line(s) containing |\Provides...|. The
% \TextUsage\FileInfo\ command takes the subsequent stuff till the
% closing |]| and subsequent line end, extracts from it the info and
% writes it to the \file{.aux} and rescans the stuff. We use an
% \eTeX\ primitive |\scantokens| for that purpose.
%
% A~macro for the standard note is provided, \TextUsage\filenote, that
% expands to ``This file has version number \<version number> dated
% \<date>.'' To place such a~note in the document's title (or heading,
% with |\DocInclude| at the default settings), there's
% \TextUsage\thfileinfo\ macro that puts |\fileinfo| in |\thanks|.
%
% \stanza
% Since |\noindent| didn't want to cooperate with my code and
% narration layers sometimes, I~provide \TextUsage\gmdnoindent\ that
% forces a~not indented paragraph if |\noindent| could not.
%
% \stanza
% If you declare the code delimiter other than |%| and then want |%|
% back, you may write \TextUsage\CDPerc\ instead of |\CodeDelim*\%|.
%
% If you like |&| as the code delimiter (as I~did twice), you may write
% \TextUsage\CDAnd\ instead of |\CodeDelim\&|.
%
% \stanza To get `\CS' which is `CS' in small caps (in |\acro| to be
% precise), you can write \TextUsage\CS. This macro is |\protected| so
% you can use it safely in |\changes| e.g. Moreover, it checks
% whether the next token is a~letter and puts a~space if so so you
% don't have to bother about \verb*|\CS\ |.
%
% \stanza
% To enumerate the list of command's arguments or macro's parameters
% there is the \TextUsage*{enumargs} environment which is a~version of
% \env{enumerate} with labels like |#7|. You can use |\item| or, at
% your option, \TextUsage\mand\ which is just an alias for the
% former. For an optional arguments use \TextUsage\opt\ which wraps
% the item label in square brackets. Moreover, to align optional and
% mandatory arguments digit under digit, use
% the \TextUsage*{enumargs*} environment.
%
% Both environments take an optional argument which is the number of
% |#|s. It's 1 by default, but also can be 2 or 4 (other numbers will
% typeset numbers without a~|#|). Please feel free to notify me if you
% really need more hashes in that environment.
%
% \stanza
% For an example driver file see chapter \gmiflink{The driver}.
%
%
% \subdivision{A~queerness of \cs{label}}
%
% You should be loyally informed that |\label| in \gmdoc\
% behaves slightly
% non-standard in the \cs{Doc\-In\-put/\:Inc\-lud\-e}d files:
% the automatic redefinitions of |\ref| at each code line
% are \emph{global} (since the code is typeset in groups and the
% |\ref|s will be out of those groups), so
% a~|\ref|erence in the narrative will point at the last code line not
% the last section, \emph{unlike} in the standard \LaTeX.
%
%
% \subdivision{\docfm-compatibility}
%
% One of my goals while writing \gmdoc\ was to make compilation of
% \docfm-like files with \gmdoc\ possible. I~cannot guarantee the goal
% has been reached but I~\emph{did} compile \pk{doc.dtx} with not
% a~smallest change of that file (actually, there was a~tiny little
% buggie in line 3299 which I~fixed remotely with
% \cs{AfterMacrocode} tool written specially for that). So, if
% you wish to compile a~\docfm-like file with my humble package, just
% try.
%
% \TextUsage\AfterMacrocode\arg{mc number}\arg{the stuff} defines
% control sequence \cs{gmd@mchook}\<mc number> with the meaning \<the\
% stuff> which is put at the end of \env{macrocode} and \env{oldmc}
% number \<mc number> (after the group).
%
% The \docfm\ commands most important in my opinion are supported by
% \gmdoc. Some commands, mostly the obsolete in my opinion, are not
% supported but give an info on the terminal and in \pk{.log}.
%
% I~assume that if one wishes to use \docfm's interface then \heshe\
% won't use \gmdoc's options but just the default. (Some \gmdoc\
% options may interfere with some \docfm\ commands, they may cancel
% them e.g.)
%
% \stanza
% The main input commands compatible with \docfm\ are
% \TextUsage\OldDocInput\ and \TextUsage\DocInclude, the latter
% however only in the \TextUsage\olddocIncludes\ declaration's scope.
%
% Within their scope/argument the \TextUsage*{macrocode} environments
% behave as in \docfm, i.e.\ they are a~kind of verbatim and require to be
% ended with |% \end{macrocode(*)}|.
%
% The default behaviour of \env{macrocode(*)} with the `new' input
% commands is different however. Remember that in the `new' fashion the code
% and narration layers philosophy is in force and that is sustained
% within \env{macrocode(*)}. Which means basically that with `new'
% settings when you write
%\begin{verbatim}
%% \begin{macrocode}
% \alittlemacro % change it to \blaargh
%%\end{macrocode}
%\end{verbatim}
% and |\blaargh|'s definition is |{foo}|, you'll get
%\[|\alittlemacro %| change it to foo\]
% (Note that `my' \env{macrocode} doesn't require
% the magical \verb*|% \end|.)
%
% If you are used to the traditional (\docfm's) \env{macrocode} and still wish to
% use \gmdoc\ new way, you have at least two options: there is the
% \TextUsage*{oldmc} environment analogous to the traditional (\docfm's)
% \env{macrocode} (it also has the starred version), that's the first
% option (I~needed the traditional behaviour once in this
% documentation, find out where \& why). The other is to write
% \TextUsage\OldMacrocodes. That declaration (\acro{OCSR}) redefines
% \env{macrocode} and \env{macrocode*} to behave the traditional way.
% (It's always executed by |\OldDocInput| and |\olddocIncludes|.)
%
%
% \stanza
% For a~more detailed discussion of what is \docfm-compatible and how,
% see the code section \gmiflink[doccompa]{\docfm-compatibiliy}.
%
%^^A \docstrip
%<*package>
% \division[The driver part]{\gmhypertarget{The driver} part}
%
% In case of a~single package, such as \pk{gmutils}, a~driver part of
% the package may look as follows and you put it before
% \cs{ProvidesPackage/Class}.
%
%\begin{verbatim}
%% \skiplines we skip the driver
%\ifnum\catcode`\@=12
%
%\documentclass[outeroff, pagella, fontspec=quiet]{gmdocc}
%\usepackage{eufrak}% for |\continuum| in the commentary.
%\twocoltoc
%\begin{document}
%
%\DocInput{\jobname.sty}
%\PrintChanges
%\thispagestyle{empty}
%\typeout{%
% Produce change log with^^J%
% makeindex -r -s gmglo.ist -o \jobname.gls \jobname.glo^^J
% (gmglo.ist should be put into some texmf/makeindex directory.)^^J}
%\typeout{%
% Produce index with^^J%
% makeindex -r \jobname^^J}
%\afterfi{\end{document}}
%
%\fi% of driver pass
%%\endskiplines
%\end{verbatim}
%
% The advantage of \TextUsage\skiplines…\TextUsage\endskiplines\ over
% |\iffalse…\fi| is that the latter has to contain balanced \cs{if}s
% and \cs{fi}s while the former hasn't because it sanitizes the
% stuff. More precisely, it uses the \cs{dospecials} list, so it
% sanitizes also the braces.
%
% Moreover, when the |countalllines(*)| option is in force,
% |\skipfiles|…|\endskipfiles| keeps the score of skipped lines.
%
% \label{Driver}
% {\def\CommonEntryCmd{UsgEntry}%
%
% Note |%\iffalse| \dots\ |%\fi| in the code layer that protects the
% driver against being typeset.
%
% But \pk{gmdoc} is more baroque and we want to see the driver
% typeset---behold.
\ifnum\catcode`\@=12
% \CodeUsgIndex*{outeroff} \CodeUsgIndex*{mwrep}
\documentclass[countalllines, codespacesgrey, outeroff, debug, mwrep,
pagella, fontspec=quiet]{gmdocc}
\twocoltoc
\title{The \pk{gmdoc} Package\\ i.e., \pk{gmdoc.sty} and
\pk{gmdocc.cls}}
\author{Grzegorz `Natror' Murzynowski}
\date{\ifcase\month\relax\or January\or February\or March\or April\or May\or
June\or July\or August\or September\or October\or November\or
December\fi\ 2008}
%|%\includeonly{gmoldcomm}|
%^^A\includeonly{gmeometric,gmoldcomm}
\begin{document}
%\skiplines
\emptify\udigits % because orig. def. raises an error `\cs{@secondoftwo} has
% an extra \}'
%\endskiplines
\maketitle
\setcounter{page}{2}% \pk{hyperref} cries if it sees two pages
% numbered~1.
\tableofcontents
\DoIndex\maketitle
%^^A\AtBegInputOnce{\showthe\catcode`\^^B}
\SelfInclude
% \label{SelfIncludeUsg}
\DocInclude{gmdocc}
% For your convenience I~decided to add the documentations of the
% three auxiliary packages:
\skipgmlonely[\stanza The remarks about installation and compiling
of the documentation are analogous to those in the chapter
\pk{gmdoc.sty} and therefore ommitted.\stanza]
\DocInclude{gmutils}
\DocInclude{gmiflink}
\DocInclude{gmverb}
\DocInclude{gmeometric}
\DocInclude{gmoldcomm}
\typeout{%
Produce change log with^^J%
makeindex -r -s gmglo.ist -o \jobname.gls \jobname.glo^^J
(gmglo.ist should be put into some texmf/makeindex directory.)^^J}
\PrintChanges
\typeout{%
Produce index with^^J%
makeindex -r \jobname^^J}
\PrintIndex
\afterfi{%
\end{document}
% MakeIndex shell commands:
makeindex -r gmdoc
makeindex -r -s gmglo.ist -o gmdocDoc.gls gmdocDoc.glo
% (\pk{gmglo.ist} should be put into some \pk{texmf/makeindex}
% directory.)
%^^A(
% And ``That's all, folks'' ;-)\,.
}\fi% of |\ifnum\catcode`\@=12|, of the driver that is.
% }^^A~of special \cs{CommonEntryCmd}'s definition.
%
% \StopEventually{\NoEOF}
%
%
% \division{The code}
%
% For debug
\catcode`\^^C=9\relax
% We set the |\catcode| of this char to \catactive\ in the comment
% layer.
% \catcode`\^^C\active
%
% \changes[^^C]{v0.98g}{06/10/10}{recatcoded for debug: in the working
% pass it's made ignored and in the documenting pass it's made active}
%
%
% \DoIndex{\par \@@par}
%
% The basic idea of this package is to re|\catcode| |^^M| (the line
% end char) and |%| (or any other comment char) so that they start and
% finish typesetting of what's between them as the \TeX\ code i.e.,
% verbatim and with the bells and whistles.
%
% The bells and whistles are (optional) numbering of the codelines,
% and automatic indexing the \CSs, possibly with
% special format for the `def' and `usage' entries.
%
% As mentioned in the preface, this package aims at a~minimal markup
% of the working code. A~package author writes \hisher\ splendid code
% and adds a~brilliant comment in |%|ed lines and that's all. Of
% course, if \heshe\ wants tomake a~|\section| or |\emph|asise,
% \heshe\ has to type respective \CSs.
%
% I~see the feature described above to be quite a~convenience,
% however it has some price. See section
% \gmiflink[afc]{Life among queer \acro{EOL}s} for details,
% here I~state only that in my opinion the price is not very high.
% \stanza
%
% More detailedly, the idea is to make |^^M| (end of line char) active
% and to define it to check if the next char i.e., the beginnig of the
% next line is a~|%| and if so
% to gobble it and just continue usual typesetting or else to start
% a~verbatim scope. In fact, every such a~line end starts a~verbatim
% scope which is immediately closed, if the next line begins with
% (leading spaces and) the code delimiter.
%
% Further details are typographical parameters of verbatim scope and
% how to restore normal settings after such a~scope so that a~code
% line could be commented and still displayed, how to deal with
% leading spaces, how to allow breaking a~moving argument in two lines
% in the comment layer, how to index and marginpar macros etc.
%
%
%\subdivision{The package options}
%
\RequirePackage{gmutils}[2008/08/30]% includes redefinition of
% \incs{newif} to make the switches \incs{protected}.
\RequirePackage{xkeyval}% we need key-vals later, but maybe we'll make
% the option key-val as well.
%
% Maybe someone wants the code lines not to be numbered.
\newif\if@linesnotnum
\DeclareOption{linesnotnum}{\@linesnotnumtrue}
% And maybe he or she wishes to declare resetting the line counter
% along with some sectioning counter him/herself.
\newif\if@uresetlinecount
\DeclareOption{uresetlinecount}{\@uresetlinecounttrue}
% And let the user be given a~possibility to count the comment
% lines.
\newif\if@countalllines
\newif\if@printalllinenos
\DeclareOption{countalllines}{%
\@countalllinestrue
\@printalllinenosfalse}
\DeclareOption{countalllines*}{%
\@countalllinestrue
\@printalllinenostrue}
% Unlike in \docfm , indexing the macros is the default and the
% default reference is the code line number.
\newif\if@noindex
\DeclareOption{noindex}{\@noindextrue}
\newif\if@pageindex
\DeclareOption{pageindex}{\@pageindextrue}
% It would be a~great honour to me if someone would like to document
% \LaTeX\ source with this humble package but I~don't think it's
% really probable so let's make an option that'll switch index exclude
% list properly (see sec.\ \gmiflink{Index exclude list}).
\newif\if@indexallmacros
\DeclareOption{indexallmacros}{\@indexallmacrostrue}
% Some document classes don't support marginpars or disable them by
% default (as my favourite Marcin Woli\nacute ski's classes).
%
% \changes{v0.98e}{06/09/23}{wrapped in \cs{@ifundefined} (a~bug fix:
% before the bare \cs{newif} falsified the \cs{if@marginparsused}
% switch when it had been defined and set True by the \pk{mwart}
% class.)}
\@ifundefined{if@marginparsused}{\newif\if@marginparsused}{}
% This switch is copied from \pk{mwbk.cls} for compatibility with
% it. Thanks to it loading an \pk{mwcls} with |[withmarginpar]| option
% shall switch marginpars on in this package, too.
%
% To be compatible with the standard classes, let's |\let|:
\@ifclassloaded{article}{\@marginparsusedtrue}{}
%\label{mparclause1}
\@ifclassloaded{report}{\@marginparsusedtrue}{}
\@ifclassloaded{book}{\@marginparsusedtrue}{}
% \label{mparclause2} And if you don't use \pk{mwcls} nor standard
% classes, then you have the options:
\DeclareOption{withmarginpar}{\@marginparsusedtrue}
\DeclareOption{nomarginpar}{\@marginparsusedfalse}
% The order of the above conditional switches and options is significant.
% Thanks to it the options are available also in the
% standard classes and in \pk{mwcls}.
%
% \stanza
% To make the code spaces blank (they are visible by default except
% the leading ones).
\DeclareOption{codespacesblank}{%
\AtEndOfPackage{% to allow \inverb|codespacesgrey, codespacesblank|
\AtBeginDocument{\CodeSpacesBlank}}}
\DeclareOption{codespacesgrey}{%
% \changes{v0.99l}{2008/08/06}{added due to Will Robertson's
% suggestion}
\AtEndOfPackage{% to put the declaration into the begin-document
% hook after definition of \incs{visiblespace}.
\AtBeginDocument{\CodeSpacesGrey}}}
\ProcessOptions
% \subdivision{The dependencies and preliminaries}
%
% We require another package of mine that provides some tricky macros
% analogous to the \LaTeX\ standard ones, such as |\newgif| and
% |\@ifnextcat|. Since 2008/08/08 it also makes \cs{if…} switches
% \cs{protected} (redefines \cs{newif})
\RequirePackage{gmutils}[2008/08/08]
% A~standard package for defining colours,
\RequirePackage{xcolor}
% and a~colour definition for the hyperlinks not to be too bright
\definecolor{deepblue}{rgb}{0,0,.85}
% And the standard package probably most important for \gmdoc:
% If the user doesn't load \pk{hyperref} with \hisher\ favourite
% options, we do, with \emph{ours}. If \heshe\ has
% done it, we change only the links' colour.
%
% \changes[hyperref]{v0.97}{06/09/04}{linkcolor made deep blue}
% \changes[hyperref]{v0.99g}{2007/10/30}{added bypass of encoding for loading
% \pk{url}}
%^^A\@ifXeTeX{\XeTeXdefaultencoding "cp1250"}{}
\@ifpackageloaded{hyperref}{\hypersetup{colorlinks=true,
linkcolor=deepblue, urlcolor=blue, filecolor=blue}}{%
\RequirePackage[colorlinks=true, linkcolor=deepblue, urlcolor=blue,
filecolor=blue, pdfstartview=FitH, pdfview=FitBH,
%^^A \@ifXeTeX{xetex,}{}
pdfpagemode=UseNone]{hyperref}}
%^^A\@ifXeTeX{\XeTeXdefaultencoding "utf-8"}{}
% \changes[hyperref]{v0.99k}{2008/08/04}{removed some lines testing if
% \XeTeX\ colliding with \pk{tikz} and most probably obsolete}
% Now a~little addition to \pk{hyperref}, a~conditional hyperlinking
% possibility with the |\gmhypertarget| and |\gmiflink| macros. It
% \emph{has} to be loaded \emph{after} \pk{hyperref}.
\RequirePackage{gmiflink}
% And a~slight redefinition of \env{verbatim}, |\verb(*)| and
% providing of |\MakeShortVerb(*)|.
\RequirePackage{gmverb}[2008/08/20]
\if@noindex
\AtBeginDocument{\gag@index}% for the latter macro see line
% \ref{gag@index}.
\else
\RequirePackage{makeidx}\makeindex
\fi
% Now, a~crucial statement about the code delimiter in the input file.
% Providing a~special declaration for the assignment is intended for
% documenting the packages that play with |%|'s
% |\catcode|. Some macros for such plays are defined
% \gmiflink[CDAnd]{further}.
%
% The declaration comes in the starred and unstarred version. The
% unstarred version besides declaring the code delimiter declares
% the same char as the verb(atim) `hyphen'. The starred version
% doesn't change the verb 'hyphen'. That is intended for the special
% tricks e.g.\ for the \env{oldmc} environment.
%
% If you want to change the verb `hyphen', there is the
% |\VerbHyphen\|\<one char>
% declaration provided by \pk{gmverb}.
%
% \changes{v0.98c}{06/9/8}{\cs{CodeDelim} renamed from a~rather confusing
% \cs{NewCommentChar}; also the internal \cs{code@delim} is renamed
% from as much confusing \cs{commentchar} and a~\cs{glet} for
% \cs{verb}s \cs{verbhyphen} is added; similarly
% \cs{[Un]CodeDelimAnd} renamed from \cs{[un]commentand}}
%
% \changes{v0.98m}{06/11/10}{\cs{CDAnd} and \cs{CDPerc} renamed from
% \cs{CodeDelimAnd} and \cs{UnCodeDelimAnd} and both those commands
% simplified to just declaring the code delimiter}
%
% \changes{v0.99a}{06/12/1}{split into the starred and unstarred
% versions}
%
\def\CodeDelim{\@ifstar\Code@Delim@St\Code@Delim}
\def\Code@Delim#1{%
{\escapechar\m@ne
\@xa\gdef\@xa\code@delim\@xa{\string#1}}}
% (|\@xa| is |\expandafter|, see \pk{gmutils}.)\DoNotIndex\@xa
\def\Code@Delim@St#1{\Code@Delim{#1}\VerbHyphen{#1}}
% It is an invariant of \gmdoc ing that |\code@delim| stores the
% current code delimiter (of catcode 12).
%
% The |\code@delim| should be \catother\ so a~space is not allowed as
% a~code delimiter. I~don't think it \emph{really} to be a~limitation.
%
% And let's assume you do as we all do:
\CodeDelim*\%
% We'll play with |\everypar|, a~bit, and if you use such things as
% the |{itemize}| environment, an error would occur if we didn't
% store the previous value of |\everypar| and didn't restore it at
% return to the narration. So let's assign a~|\toks| list to store the
% original |\everypar|:
\newtoks\gmd@preverypar
\newcommand*\settexcodehangi{%
\hangindent=\verbatimhangindent \hangafter=\@ne}% we'll use
% it in the inline comment case. |\verbatimhangindent| is provided by the
% \pk{gmverb} package and\equals3\,em by default.
% \DefIndex\@@settexcodehangi
\@ifdefinable\@@settexcodehangi{\let\@@settexcodehangi=\settexcodehangi}
%^^A~\gmdeverycodeline{\@@settexcodehangi}
% We'll play a~bit with |\leftskip|, so let the user have a~parameter
% instead. For normal text (i.e.\ the comment):
\newlength\TextIndent
% I~assume it's originally equal to |\leftskip|, i.e.\ |\z@|. And for
% the \TeX\ code:
% \HideAllDefining
\newlength\CodeIndent
% \label{CodeIndent}
% \Define\CodeIndent
\CodeIndent=1,5em\relax
% And the vertical space to be inserted where there are blank lines in the
% source code:
\@ifundefined{stanzaskip}{\newlength\stanzaskip}{}
% I~use |\stanzaskip| in \pk{gmverse} package and
% derivatives for typesetting poetry. A~computer program code \emph{is}
% poetry.
% \Define\stanzaskip
\stanzaskip=\medskipamount
\advance\stanzaskip by-.25\medskipamount% to preserve the stretch- and
% shrinkability.\par
%
%\stanza
% A~vertical space between the commentary and the code seems
% to enhance readability so declare
\newskip\CodeTopsep
\newskip\MacroTopsep
% And let's set them. For \ae sthetic
% minimality\StraightEOL\footnote{The terms `minimal' and `minimalist'
% used in \gmdoc\ are among others\ inspired by the \emph{South Park}
% cartoon's episode \emph{Mr.\ Hankey The~Christmas (\dots)}\/ in
% which `Philip Glass, a~Minimalist New York composer' appears in
% a~`non-denominational non-offensive Christmas play' ^^A(
% ;-)\,. (Philip Glass composed the music to the \emph{Qatsi}
% trilogy among others).}\QueerEOL\ let's unify them and the other most important
% vertical spaces used in \gmdoc. I~think a~macro that gathers all
% these assignments may be handy.
% \ResumeAllDefining
% \Define\MacroTopsep
% \Define\CodeTopsep
% \Define\UniformSkips
\def\UniformSkips{%\label{UniformSkips}
% \Define\CodeTopsep \Define\MacroTopsep
\CodeTopsep=\stanzaskip
\MacroTopsep=\stanzaskip
\abovedisplayskip=\stanzaskip
%\nostanza \leftskip0sp \gmdnoindent
% |%% \abovedisplayshortskip|
% remains untouched as it is 0.0\,pt plus 3.0\,pt by default.
% \nostanza
\belowdisplayskip=\stanzaskip
\belowdisplayshortskip=.5\stanzaskip% due to DEK's idea of making the
% short below display skip half of the normal.
\advance\belowdisplayshortskip by\smallskipamount
\advance\belowdisplayshortskip by-1\smallskipamount% We advance
% \incs{be\+low\+dis\+play\+short\+skip} forth and back to
% give it the \nlpercent\cs{small\+skip\+am\+ount}'s shrink- and
% stretchability components.
\topsep=\stanzaskip
\partopsep=\z@
}
% We make it the default,
\UniformSkips
% \gmdnoindent but we allow you to change the benchmark glue
% i.e., |\stanzaskip| in the preamble and still have the other glues
% set due to it: we launch |\UniformSkips| again after the
% preamble.
\AtBeginDocument{\UniformSkips}
% So, if you don't want them at all i.e., you don't want to set other
% glues due to |\stanzaskip|, you should use the following
% declaration. That shall discard the unwanted setting already placed
% in the |\begin{document}| hook.
%
\newcommand*\NonUniformSkips{\@relaxen\UniformSkips}
% Why do we launch |\UniformSkips| twice then? The first time is
% to set all the \gmdoc-specific glues \emph{somehow}, which allows
% you to set not all of them, and the second
% time to set them due to a~possible change of |\stanzaskip|.
%
% \stanza
% And let's define a~macro to insert a~space
% for a~chunk of documentation, e.g., to mark the beginning of new
% macro's explanation and code.
\newcommand*\chunkskip{%
\par\addvspace{%
\glueexpr\MacroTopsep
\if@codeskipput-\CodeTopsep\fi
\relax
}\@codeskipputgtrue}
% And, for a~smaller part of text,
\pdef\stanza{%
\par\addvspace{%
\glueexpr\stanzaskip
\if@codeskipput-\CodeTopsep\fi
\relax}\@codeskipputgtrue}
% Since the stanza skips are inserted automatically most often (cf.\ lines
% \ref{codeskip}, \ref{codeskip2}, \ref{codeskip3}, \ref{codeskip4},
% \ref{codeskip5}), sometimes you may need to forbid them.
\newcommand*\nostanza{% \changes{v0.99n}{2008/08/21}{added adding
% negative skip if in vmode and \cs{par}}
\par
\if@codeskipput\unless\if@nostanza\vskip-\CodeTopsep\relax\fi\fi
\@codeskipputgtrue\@nostanzagtrue
\@afternarrgfalse\@aftercodegtrue}% In the `code
% to narration' case the first switch is enough but in the countercase
% `narration to code' both the second and third are necessary while
% the first is not.
%
%\stanza
% To count the lines where they have begun not before them\nostanza
\newgif\if@newline
% |\newgif| is |\newif| with global effect i.e., it defines |\...gtrue|
% and |\...gfalse| switchers that switch respective Boolean switch
% \emph{globally}. See \pk{gmutils} package for details.
% \stanza
%
% To handle the \ds\ directives not \emph{any} |%<...|\,.
% \Define\if@dsdir
\newgif\if@dsdir
% This switch will be falsified at the first char of a~code line. (We
% need a~switch independent of the one indicationg whether the line
% has or has not been counted because of two reasons: 1.\ line
% numbering is optional, 2.\ counting the line falsifies that switch
% \emph{before} the first char.)
%
% \subdivision{The core}
%
% Now we define main |\input|ing command that'll change catcodes.
% The macros used by it are defined later.
%
\newcommand*\DocInput{\bgroup\@makeother\_\Doc@Input}
\begingroup\catcode`\^^M=\active%
\firstofone{\endgroup%
\newcommand*{\Doc@Input}[1]{\egroup\begingroup%
% \changes{v0.98}{06/09/05}{\cs{@makeother\protect\bslash_} added}
% \DefIndex\gmd@inputname
\edef\gmd@inputname{#1}% we'll use it in some notifications.
% \DefIndex\gmd@currentlabel@before
\let\gmd@currentlabel@before=\@currentlabel% we store it because
% we'll do |\xdef|s of |\@currentlabel| to make proper references
% to the line numbers so we want to restore current
% |\@currentlabel| after our group.\
% \CodeUsgIndex\clubpenalty \CodeUsgIndex\widowpenalty
\gmd@setclubpenalty% we wrapped the assignment of |\clubpenalty|
% in a~macro because we'll repeat it twice more.
\@clubpenalty\clubpenalty \widowpenalty=3333 % Most paragraphs of the code will be
% one-line most probably and many of the narration, too.\par\
% \changes[\DocInput]{v0.95}{06/08/15}{\cs{club-}
% and \cs{widowpenalty} set both to 3333}
%\CodeUsgIndex\tolerance
\tolerance=1000 % as in \docfm.
% ^^A \catcode`\^^M=\active% redundant with line 2304
% \CodeUsgIndex\code@delim
\@xa\@makeother\csname\code@delim\endcsname%
% \CodeUsgIndex\gmd@resetlinecount
\gmd@resetlinecount% due to the option |uresetlinecount|
% we reset the linenumber counter or do nothing.
% \Define*{^^M}
\QueerEOL% \changes{v0.99m}{2008/08/09}{there was
% \cs{let}\hathat{M} but \cs{QueerEOL} is better: it also
% redefines \cs{}\hathat M} It has to be before the
% begin-input-hook to allow change by that hook.
% \CodeUsgIndex\@beginputhook
\@beginputhook% my first use of it is to redefine |\maketitle|
% just at this point not globally. \CodeUsgIndex\everypar
\everypar=\@xa{\@xa\@codetonarrskip\the\everypar}%
% \Define\gmd@guardedinput
\edef\gmd@guardedinput{%
\@nx\@@input #1\relax% |\@nx| is
% |\noexpand|, see \pk{gmutils}.\DoNotIndex\@nx
% \incs{@@input} is the
% true \TeX's \incs{input}. \CodeUsgIndex\EOFMark
\gmd@iihook% cf.\ line \ref{iihook}
\@nx\EOFMark% to pretty finish the input, see
% line~\ref{eofMark}.\CodeUsgIndex\code@delim
\@nx\CodeDelim\@xa\@nx\csname\code@delim\endcsname% to
% ensure the code delimiter is the same as at the beginning of
% input.
% \changes[\DocInput]{v0.99c}{2007/03/02}{added ensuring the
% code delimiter to be the same at the end as at the beginning}
\@nx^^M\code@delim%
% \label{guardians}
}% we add guardians after
% |\input|ing a~file; somehow an error occurred without them.
\catcode`\%=9 % for \docfm -compatibility.\label{ignorePercent}
\setcounter{CheckSum}{0}% we initialize the counter for the number
% of the escape chars (the assignment is |\global|).
\everyeof{\relax}% |\@nx| moved not to spoil input of toc e.g.
\@xa\@xa\@xa^^M\gmd@guardedinput%\label{eeeEOL}
\par%
% \CodeUsgIndex\@endinputhook
\@endinputhook% It's a~hook to let postpone
% some stuff till the end of input. We use it e.g.\ for the
% \docfm-(not)likeliness notifications.
\glet\@currentlabel=\gmd@currentlabel@before% we restore
% value from before this group. In a~very special case this could
% cause unexpected behaviour of crossrefs, but anyway we acted
% globally and so acts \pk{hyperref}.
\endgroup%
}% end of |\Doc@Input|'s definition.
}% end of |\firstofone|'s argument.
% So, having the main macro outlined, let's fill in the details.
%
% First, define the queer \acro{EOL}. We define a~macro that |^^M| will be
% let to. |\gmd@textEOL|
% will be used also for checking the |%^^M| case (|\@ifnextchar|
% does |\ifx|).
%
\pdef\gmd@textEOL{ % a~space just like in normal
% \TeX. We put it first to % cooperate with \inverb|\^^M|'s
% |\expandafter\ignorespaces|. It's % no % problem since a~space
% | |${}_{10}$ doesn't drive \TeX\ out of % the vmode.
\ifhmode\@afternarrgtrue\@codeskipputgfalse\fi% being in
% the horizontal mode means we've just typeset some narration so we
% turn the respective switches: the one bringing the message `we are
% after narration' to True (|@afternarr|) and the `we have put the
% code-narration glue' to False (|@codeskipput|). Since we are in
% a~verbatim group and the information should be brought outside it,
% we switch the switches globally (the letter |g| in both).
\@newlinegtrue% to |\refstep| the lines' counter at the proper
% point.
\@dsdirgtrue% to handle the \ds\ directives.
\@xa\@trimandstore\the\everypar\@trimandstore%we store the
% previous value of |\everypar| register to restore it at
% a~proper point. See line \ref{@trimandstore} for the details.
\begingroup%^^A ##### maybe we prefer to restore the genuine |\par|
% ^^A~before the group?
% ^^A \let\gmd@narrpar\par% we store the possibly redefined |\par| to
% ^^A~typeset the previous narration appropriately.
% ^^A \def\par{\gmd@narrpar\let\par\@@par}% the first |\par| will be
% ^^A~the one brought from the narration but then,
\gmd@setclubpenalty% Most paragraphs will be
% one-line most probably. Since some sectioning commands may change
% \incs{clubpenalty}, we set it again here and also after this
% group.
\aftergroup\gmd@setclubpenalty%
\let\par\@@par% inside the verbatim group we wish |\par| to be genuine.
% \CodeUsgIndex\ttverbatim
\ttverbatim% it does |\tt| and makes specials
% other or |\active|-and-breakable.
\gmd@DoTeXCodeSpace%
\@makeother\|% because |\ttverbatim| doesn't do that.
\MakePrivateLetters% see line \ref{MPL}.\par
\@xa\@makeother\code@delim% we are
% almost sure the code comment char is among the chars having
% been \catother ed already. For `almost' see the
% \gmiflink[IndexInput]{\cs{IndexInput}} macro's definition.\par
% So, we've opened a~verbatim group and want to peek at the next
% character. If it's |%|, then we just continue narration, else we
% process the leading spaces supposed there are any and, if after
% them is a~|%|, we just continue the commentary as in the
% previous case or else we typeset the \TeX\ code.
\@xa\@ifnextchar\@xa{\code@delim}{%\label{ifContNarr}
% \CodeUsgIndex\gmd@continuenarration
\gmd@continuenarration}{% \CodeUsgIndex\gmd@dolspaces
\gmd@dolspaces% it will launch |\gmd@typesettexcode|.
}% end of |\@ifnextchar|'s else.
}% end of |\gmd@textEOL|'s definition.
\def\gmd@setclubpenalty{\clubpenalty=3333 }
% \stanza
% For convenient adding things to the begin- and endinput hooks:
\def\AtEndInput{\g@addto@macro\@endinputhook}
\def\@endinputhook{}
%\label{@endinputhook}
% Simili modo
\def\AtBegInput{\g@addto@macro\@beginputhook}
\def\@beginputhook{}
% For the index input hooking now declare a~macro, we define it
% another way at line \ref{iihook}.
\emptify\gmd@iihook
% And let's use it instantly to avoid a~disaster while reading in the
% table of contents.
% \DefIndex\gmd@@toc
\AtBegInput{\let\gmd@@toc\tableofcontents
\def\tableofcontents{% \label{straighttoc}
\@ifQueerEOL{\StraightEOL\gmd@@toc\QueerEOL}%
{\gmd@@toc}}}
% As you'll learn from lines \ref{QueerEOL} and \ref{StraightEOL}, we
% use those two strange declarations to change and restore the very
% special meaning of the line end. Without such changes
% |\tableofcontents| would cause a~disaster (it did indeed). And to
% check the catcode of |^^M| is the r\ocircum le of |\@ifEOLactive|:
%
% \changes{v0.98a}{06/09/06}{\cs{tableofcontents}, \cs{PrintIndex}
% and \cs{PrintChanges} modified to work properly in the `queer line
% ends' environment.}
\long\def\@ifEOLactive#1#2{%
\ifnum\catcode`\^^M=\active \afterfi{#1}\else\afterfi{#2}\fi}
\foone\obeylines{%
\long\def\@ifQueerEOL#1#2{%% \changes{v0.98a}{06/09/06}{added}
\@ifEOLactive{\ifx^^M\gmd@textEOL\afterfi{#1}\else\afterfi{#2}\fi}%
{#2}}% of \cs{@ifQueerEOL}
}% of \cs{foone}
% The declaration below is useful if you
% wish to put sth.\ just in the nearest input/included file and no
% else: at the moment of putting the stuff it will erase it from the
% hook. You may declare several |\AtBegInputOnce|s, they add up.
%
%^^A Define*{hook@oncecnt}
%^^A CodeDefIndex*\c@hook@oncecnt
%^^A\newcounter{hook@oncecnt}
% \Define\gmd@ABIOnce
\@emptify\gmd@ABIOnce
\AtBegInput\gmd@ABIOnce
\long\def\AtBegInputOnce#1{%
% \changes{v0.98a}{06/09/05}{The counter \cs[]{hook@oncecnt} added
% to number the self-destructing cs's (thence the hooks add up) and
% the prefix changed to \cs[]{gmd/AtBI/NeuroOncer}}
% \changes{v0.98l}{06/10/26}{After the Tiger's suggestion, defining
% a~unique macro for each use of \* substituted with just one macro in
% the begin input hook and adding to this macro.}
%^^A \stepcounter{hook@oncecnt}%
%^^A \@xa\long\@xa\edef% \CodeUsgIndex*{NeuroOncer}
%^^A \csname gmd/AtBI/NeuroOncer\the\c@hook@oncecnt\endcsname{%
%^^A \@nx\glet\csname gmd/AtBI/NeuroOncer\the\c@hook@oncecnt\endcsname\relax}%
%^^A \@xa\addtomacro\csname gmd/AtBI/NeuroOncer\the\c@hook@oncecnt\endcsname{#1}%
%^^A \@xa\AtBegInput\csname gmd/AtBI/NeuroOncer\the\c@hook@oncecnt\endcsname
\gaddtomacro\gmd@ABIOnce{\g@emptify\gmd@ABIOnce#1}}
% Many tries of finishing the input cleanly led me to setting the
% guardians as in line \ref{guardians} and to
\def\EOFMark{\<eof>}
% \label{eofMark}Other solutions did print the last code delimiter or
% would require managing a~special case for the macros typesetting
% \TeX\ code to suppress the last line's numbering etc.
%
% If you don't like it, see line \ref{NoEOF}.
% \stanza
%
% Due to the |codespacesblank| option in the line \ref{cspb} we launch
% the macro defined below to change the meaning of a~\pk{gmdoc}-kernel
% macro.
% \DefIndex\gmd@DoTeXCodeSpace
\begin{obeyspaces}%
\gdef\CodeSpacesVisible{%
\def\gmd@DoTeXCodeSpace{%
\obeyspaces\let =\breakablevisspace}}%
% \stanza
% \Define\CodeSpacesBlank
% \changes{v0.98a}{06/09/06}{a~kernel \cs{gmd@verbatimspace} renamed
% to \cs{gmd@texcodespace} and a~bug fixed.}
% \CodeUsgIndex\gmobeyspaces
% \DefIndex\gmd@texcodespace
\gdef\CodeSpacesBlank{%
\let\gmd@DoTeXCodeSpace\gmobeyspaces%
\let\gmd@texcodespace=\ }% the latter |\let| is for the |\if...|s.
%\stanza
% \Define\CodeSpacesSmall
\gdef\CodeSpacesSmall{%
\def\gmd@DoTeXCodeSpace{%
\obeyspaces\def {\,\hskip\z@}}%
\def\gmd@texcodespace{\,\hskip\z@}}%
%\stanza
\end{obeyspaces}
\def\CodeSpacesGrey{%
% \changes{v0.99l}{2008/08/06}{added due to Will Robertson's
% suggestion}
\CodeSpacesVisible
\VisSpacesGrey% defined in \pk{gmverb}
}%
% Note that \cs{CodeSpacesVisible} doesn't revert \cs{CodeSpacesGrey}.
\CodeSpacesVisible
% How the continuing of the narration should look like?
\def\gmd@continuenarration{%
\endgroup
\gmd@cpnarrline% see \gmiflink[countnarr]{below}.
\@xa\@trimandstore\the\everypar\@trimandstore
\everypar=\@xa{\@xa\@codetonarrskip\the\everypar}%
\@xa\gmd@checkifEOL\@gobble}
% Simple, isn't it? (We gobble the `other' code delimiter. Despite of
% |\egroup| it's \catother\ because it was touched by |\futurelet|
% contained in |\@ifnextchar| in line \ref{ifContNarr}. And in line
% \ref{ContNarr2} it's been read as \catother. That's why it
% works in spite of that |%| is of category `ignored'.)
\if@countalllines
%^^A erased obsolete code is backed up in gmdoc.sty~5~.
%\gmhypertarget[countnarr]{If the} |countalllines| option is in force,
% we get the count of lines from the \cs{inputlineno} primitive. But
% if the option is |countalllines*|, we want to print the line number.
%^^A \def\gmd@countnarrline{%
%^^A \if@newline
%^^A \gmd@countnarrline@
%^^A \fi}% \DefIndex\gmd@countnarrline
\def\gmd@countnarrline@{%
\gmd@grefstep{codelinenum}\@newlinegfalse
\everypar=\@xa{%
\@xa\@codetonarrskip\the\gmd@preverypar}%^^A
% the \cs{hy\+p\+er\+lab\+el\+@\+li\+ne} macro puts
% a~hypertarget in a~|\raise| i.e., drives \TeX\ into the
% horizontal mode so |\everypar| shall be issued. Therefore we
% should restore it.
}% of \cs{gmd@countnarrline@}
\def\gmd@grefstep#1{% instead of diligent redefining all possible
% commands and environments we just assign the current value of
% the respective \TeX's primitive to the \env{codelinenum}
% counter. Note we decrease it by $-1$ to get the proper value
% for the next line. (Well, I~don't quite know why, but it
% works.) %^^A~we'll step its value by 1 because it will be put
% ^^Aat the beninning of the \emph{next} line.
\ifnum\value{#1}<\inputlineno
\csname c@#1\endcsname\numexpr\inputlineno-1\relax
\ifvmode\leavevmode\fi% this line is added 2008/08/10 after an
% ^^A(
% all-night debuggery ;-) that showed that at one point
% \incs{gmd@grefstep} was called in vmode which caused adding
% \incs{penalty 10000} to the main vertical list and thus
% forbidding pagebreak during entire \inenv{oldmc}.
\grefstepcounter{#1}%
\fi}% We wrap stepping the counter in an \incs{ifnum} to avoid
% repetition of the same ref-value (what would result in the
% ``multiply defined labels'' warning).\par\
% The
% \cs{grefstepcounter} macro, defined in \pk{gmverb}, is
% a~global
% version of \cs{ref\-step\-count\-er}, observing the redefinition
% made
% to \incs{refstepcounter} by \pk{hyperref}.
\if@printalllinenos% Note that checking this swich makes only
% sense when |countalllines| is true.
\def\gmd@cpnarrline{% count and print narration line
\if@newline
\gmd@countnarrline@
\hyperlabel@line
{\LineNumFont\thecodelinenum}\,\ignorespaces}%
\fi}
\else% not |printalllinenos|
\emptify\gmd@cpnarrline
\fi
\def\gmd@ctallsetup{% In the \inenv{oldmc} environments and with the
% \incs{FileInfo} declaration (when % \inenv{countalllines} option
% is in force) the % code is gobbled as an argument of a~macro and
% % then processed at one place (at the end of % \inenv{oldmc}
% % e.g.) so if we used % \incs{inputlineno}, we would have got all
% % the % lines with the same number. But we only set the counter
% % not \incs{refstep} it to avoid putting a~hypertarget.
\setcounter{codelinenum}{\inputlineno}% it's global.
\let\gmd@grefstep\hgrefstepcounter}
\else% not |countallines| (and therefore we won't print the narration
% lines' numbers either)
\@emptify\gmd@cpnarrline
\let\gmd@grefstep\hgrefstepcounter% \label{let grefstep} if we don't
% want to count all the lines, we only \incs{ref}-increase the
% counter in the code layer.
\emptify\gmd@ctallsetup
\fi% of |\if@countalllines|
\def\skiplines{\bgroup
\let\do\@makeother \dospecials % not \incs{@sanitize} because the
% latter doesn't recatcode braces and we want all to be quieten.\ilrr
%^^A \catcode`\^^M\active what was it for? for countalllines, but
%^^A~there's no need of.
\gmd@skiplines}
\edef\gmu@tempa{%
\long\def\@nx\gmd@skiplines##1\bslash endskiplines{\egroup}}
\gmu@tempa
% And typesetting the \TeX\ code?
\foone\obeylines{% \DefIndex\gmd@typesettexcode
\def\gmd@typesettexcode{%
\gmd@parfixclosingspace% it's to eat a~space closing the
% paragraph, see \gmiflink[closingspace]{below}. It contains
% |\par|.
%
% ^^A\special{color push gray 0.2645}%
% A~verbatim group has already been opened by \cs{ttverb\+at\+im} and
% additional \cs{cat\-code}.
\everypar={\@@settexcodehangi}% At first attempt we thought
% of giving the user a~|\toks| list to insert at the beginning of
% every code line, but what for?
% \Define*{^^M}^^A
\def^^M{% \TeX\ code \acro{EOL}
\@newlinegtrue% to |\refstep| the counter in proper place.
\@dsdirgtrue% to handle the \ds\ directives.
\global\gmd@closingspacewd=\z@% \label{noclosingspace}we
% don't wish to eat a~closing space after a~codeline, because
% there isn't any and a~negative rigid |\hskip| added to
% |\parfillskip| would produce a~blank line.
\ifhmode\par\@codeskipputgfalse\else%
\if@codeskipput%
\else\addvspace{\stanzaskip}\@codeskipputgtrue%
\fi% if we've just met a~blank (code) line, we insert
% a~|\stanzaskip| glue.
% \label{codeskip}
\fi%
\prevhmodegfalse% we want to know later that now we are in the
% vmode.
% ^^A\special{color push gray 0.2666}%
\@ifnextchar{\gmd@texcodespace}{%
\@dsdirgfalse\gmd@dolspaces}{\gmd@charbychar}%
}% end of |^^M|'s definition.\label{debug!2}
% \DefIndex\gmd@texcodeEOL
\let\gmd@texcodeEOL=^^M% for further checks inside |\gmd@charbychar|.
\raggedright\leftskip=\CodeIndent%
\if@aftercode%
\gmd@nocodeskip1{iaC}%
\else%
\if@afternarr%
% ^^A~\def^^C{\showthe\hyphenpenalty\showthe\rightskip}
\if@codeskipput\else%
\gmd@codeskip1\@aftercodegfalse%
\fi% \label{codeskip3}
\else\gmd@nocodeskip1{naN}%
\fi%
\fi% if now we are
% switching from the narration into the code, we insert a~proper
% vertical space.
\@aftercodegtrue\@afternarrgfalse%
%^^A\special{color push gray 0.2682}% before that penalty
\ifdim\gmd@ldspaceswd>\z@% and here the leading spaces.
\leavevmode\@dsdirgfalse%
\if@newline\gmd@grefstep{codelinenum}\@newlinegfalse%
\fi%
\printlinenumber% if we don't want the lines to be numbered,
% the respective option \incs{let}s this \CS to \incs{relax}.
\hyperlabel@line%
%^^A\special{color push gray 0.2689}% seems O.K.
\mark@envir% index and/or marginize an environment if there is
% some to be done so, see line \ref{mark@envir}.
\hskip\gmd@ldspaceswd%
\advance\hangindent by\gmd@ldspaceswd%
\xdef\settexcodehangi{%
\@nx\hangindent=\the\hangindent% and also set the
% hanging indent setting for the same line comment case. \acro{BTW}.,
% this |%| or rather lack of it costed me five hours of
% debugging and rewriting. Active lineends require extreme
% caution.
\@nx\hangafter=1\space}%^^A~\@nx\relax
\else%
\glet\settexcodehangi=\@@settexcodehangi%\\
%|%| |\printlinenumber| here produced line numbers for blank lines
% which is what we don't want.
\fi% of |\ifdim|
\gmd@ldspaceswd=\z@%
\prevhmodegfalse% we have done |\par| so we are not in the
% hmode.
\@aftercodegtrue% we want to know later that now we are
% typesetting a~codeline.
\if@ilgroup\aftergroup\egroup\@ilgroupfalse\fi% when we are in the inline
% comment group (for ragged right or justified), we want
% to \label{inline.egroup.2}
% close it. But if we did it
% here, we would close the verbatim group for the code. But we set
% the swich false not to repeat \inverb|\aftergroup\egroup|.
% ^^A\special{color push gray 0.2712}% before that penalty
\gmd@charbychar% we'll eat the code char by char to scan all the
% macros and thus to deal properly with the case |\%| in which the
% |%| will be scanned and won't
% launch closing of the verbatim group.
}% of |\gmd@typesettexcode|.
}% of |\foone\obeylines|.
%
% \stanza
%
% Now let's deal with the leading spaces once forever. We wish not to
% typeset | |s but to add the width of every leading space to the
% paragraph's indent and to the hanging indent, but only if there'll
% be any code character not being |%| in this line (e.g., the end of
% line). If there'll be only |%|, we want just to continue the comment
% or start a~new one. (We don't have to worry about whether we should
% |\par| or not.)
% \DefIndex\gmd@spacewd
\newlength\gmd@spacewd% to store the width of a~(leading)
% | |\catother.
% \DefIndex\gmd@ldspaceswd
\newlength\gmd@ldspaceswd% to store total length of gobbled leading
%spaces.
% It costed me some time to reach that in my verbatim scope a~space
% isn't \catother\ but \catactive, namely |\let| to
% |\breakablevisspace|.
% So let us |\let| for future:
% \Define\gmd@texcodespace
\let\gmd@texcodespace=\breakablevisspace
% And now let's try to deal with those spaces.
\def\gmd@dolspaces{%
\ifx\gmd@texcodespace\@let@token
\@dsdirgfalse
\afterfi{\settowidth{\gmd@spacewd}{\visiblespace}%
\gmd@ldspaceswd=\z@
\gmd@eatlspace}%
\else\afterfi{% about this smart macro and other of its family see
% \pk{gmutils} sec.\,3.
% ^^A \special{color push gray 0.2748}% debug of \penalty10000
% ^^A~2008/08/10 was far
% ^^A \ifnum\inputlineno>1016 \ifnum\inputlineno<1050 \show\par\fi\fi
% ^^A~O.K.
\if@afternarr\if@aftercode
\ifilrr\bgroup \gmd@setilrr\fi
\fi\fi
\par% possibly after narration
\if@afternarr\if@aftercode
\ifilrr\egroup\fi
\fi\fi
\gmd@typesettexcode}%
\fi}
% And now, the iterating inner macro that'll eat the leading spaces.
\def\gmd@eatlspace#1{%
\ifx\gmd@texcodespace#1%
\advance\gmd@ldspaceswd by\gmd@spacewd% we don't
% \incs{advance} it \incs{global}ly because the current group may be closed
% iff we meet \inverb|%| and then we'll won't indent the line anyway.\ilrr
\afteriffifi\gmd@eatlspace
\else
\if\code@delim\@nx#1%
\gmd@ldspaceswd=\z@
\afterfifi{\gmd@continuenarration#1}%\label{ContNarr2}
% \changes{v0.99n}{2008/08/30}{\cs{afterfifi} added---a~bug fix}
\else \afterfifi{\gmd@typesettexcode#1}%
\fi
\fi}%
% We want to know whether we were in hmode before reading current
% |\code@delim|. We'll need to switch the switch globally.
\newgif\ifprevhmode
% \addvspace\medskipamount
%
% And the main iterating inner macro which eats every single char of
% verbatim text to check the end. The case |\%| should be excluded
% and it is indeed.
\newcommand*\gmd@charbychar[1]{%
\ifhmode\prevhmodegtrue
\else\prevhmodegfalse
% ^^A~\special{color push gray 0.2789}%
\fi
\if\code@delim\@nx#1%
\def\next{% occurs when next a~\cs{hskip4.875pt} is to be put
% ^^A\special{color push gray 0.2791}% O.K.
\gmd@percenthack% to typeset |%| if a~comment
% continues the~codeline.
\endgroup%
\gmd@checkifEOLmixd}% to see if next is |^^M| and then do |\par|.
\else% i.e., we've not met the code delimiter
\ifx\relax#1\def\next{%
% ^^A\special{color push gray 0.2798}%
\endgroup}% special case of end of file thanks to |\everyeof|.
\else
\if\code@escape@char\@nx#1%
\@dsdirgfalse% yes, just here not before the whole |\if| because
% then we would discard checking for \ds\ directives doable by
% the active |%| at the `old macrocode' setting.
\def\next{%
%^^A \debug@special{2808}%
\gmd@counttheline#1\scan@macro}%
\else
\def\next{%
%^^A \debug@special{2806}% before penalty
\gmd@EOLorcharbychar#1}%
\fi
\fi
\fi\next}
\def\debug@special#1{%
\ifhmode\special{color push gray 0.#1}%
\else\special{color push gray 0.#1000}\fi}
% One more inner macro because |^^M| in \TeX\ code wants to peek at the
% next char and possibly launch |\gmd@charbychar|. We deal with
% counting the lines thorougly. Increasing the counter is divided into
% cases and it's very low level in one case because |\refstepcounter| and
% |\stepcounter| added some stuff that caused blank lines, at
% least with \pk {hyperref} package loaded.
\def\gmd@EOLorcharbychar#1{%
%^^A \debug@special{2829}%
\ifx\gmd@texcodeEOL#1%
\if@newline
% ^^A\special{color push gray 0.281600}% no occurence
% ^^A \if@countalllines\global\advance\c@codelinenum by\@ne obsolete
% ^^A\fi
\@newlinegfalse
\fi
\afterfi{#1}%\label{printhashone1}here we print |#1|.
\else% i.e., |#1| is \emph{not} a~(very active) line end,
\afterfi
{%^^A \debug@special{2839}% this occurs frequently
\gmd@counttheline#1\gmd@charbychar}% \label{printhashone2}or here
% we print |#1|. Here we would also possibly mark an environment
% but there's no need of it because declaring an environment to
% be marked requires a~bit of commentary and here we are after
% a~code |^^M| with no commentary.
\fi}
\def\gmd@counttheline{%
\ifvmode
\if@newline
\leavevmode
%^^A \debug@special{2851}%
\gmd@grefstep{codelinenum}\@newlinegfalse
\hyperlabel@line
\fi
%^^A\debug@special{2853}%
\printlinenumber
% ^^A\debug@special{2855}%
\mark@envir
\else% not vmode
\if@newline
% ^^A\special{color push gray 0.2842}% didn't occur
\gmd@grefstep{codelinenum}\@newlinegfalse
\hyperlabel@line
\fi
\fi}
% If before reading current |%| char we were in horizontal mode, then
% we wish to print |%| (or another code delimiter).
\def\gmd@percenthack{%
\ifprevhmode\code@delim\aftergroup~% We add a~space after |%|,
% because I~think it looks better. It's done |\aftergroup| to make
% the spaces possible after the |%| not to be
% typeset.
% \changes{v0.99n}{2008/08/21}{\cs{space} replaced with a~tilde to
% forbid a~linebreak before an inline comment}
\else\aftergroup\gmd@narrcheckifds@ne% remember that
% \cs{gmd\-@\-pre\-cent\-hack} is only called when we've the code
% delimiter and soon we'll close the verbatim group and right after
% |\endgroup| there waits |\gmd@checkifEOLmixd|.
\fi}
\def\gmd@narrcheckifds@ne#1{%^^A\typeout{narr@ne if ds \on@line}%
\@dsdirgfalse\@ifnextchar<{%
\@xa\gmd@docstripdirective\@gobble}{#1}}
% The macro below is used to look for the |%^^M| case to make
% a~commented blank line make a~new paragraph.
% Long searched and very simple at last.
\def\gmd@checkifEOL{%
\gmd@cpnarrline
\everypar=\@xa{\@xa\@codetonarrskip% we add the
% macro that'll insert a~vertical space if we leave the code and
% enter the narration.
\the\gmd@preverypar}%
\@ifnextchar{\gmd@textEOL}{%
%^^A\@ifnextMac{%^^A}
\@dsdirgfalse
\par\ignorespaces}{%
\gmd@narrcheckifds}}%
% We check if it's |%<|, a~\ds\ directive that is.
\def\gmd@narrcheckifds{%^^A\typeout{narr if ds \on@line}%
\@dsdirgfalse\@ifnextchar<{%
\@xa\gmd@docstripdirective\@gobble}{\ignorespaces}}
% In the `mixed' line case it should be a~bit more complex, though. On
% the other hand, there's no need to checking for \ds\ directives.
\def\gmd@checkifEOLmixd{%
\gmd@cpnarrline
\everypar=\@xa{\@xa\@codetonarrskip\the\gmd@preverypar}%
%
% \label{longlinethatshouldbebroken}
\@afternarrgfalse\@aftercodegtrue
\ifhmode\@codeskipputgfalse\fi
\@ifnextchar{\gmd@textEOL}{%
%^^A\@ifnextMac{%^^A}
{\raggedright\gmd@endpe\par}% without \incs{raggedright} this
% \incs{par}
% would be justified which is not appropriate for a~long codeline
% that should be broken, e.g., \ref{longlinethatshouldbebroken}.
\prevhmodegfalse
\gmd@endpe\ignorespaces}{%
% If a~codeline ends with |%|
% (|prevhmode|${}=={}$True) first |\gmd@endpe| sets the parameters
% at the \TeX\ code values and |\par| closes a~paragraph and the
% latter |\gmd@endpe| sets the parameters at the narration values.
% In the other case both |\gmd@endpe|s do the same
% and |\par| between them does nothing.
% \DefIndex\par
\def\par{% the narration \cs{par}.
\ifhmode% (I~added this |\ifhmode| as a~result of a~heavy
% debug.)
\if@afternarr\if@aftercode
\unless\if@ilgroup\bgroup\@ilgrouptrue\fi
\ifilrr\gmd@setilrr\fi
\fi\fi
\@@par
\if@afternarr
\if@aftercode
\if@ilgroup\egroup\fi% \label{inline.egroup.1}if we are both after code
% and after narration it means we are after an inline
% comment. Then we probably end a~group opened in line
% \ref{inline.bgroup}
\if@codeskipput\else\gmd@codeskip2\@aftercodegfalse\fi
%\label{codeskip4}
\else\gmd@nocodeskip2{naC}%
\fi
\else\gmd@nocodeskip2{naN}%
\fi
\prevhmodegfalse\gmd@endpe% when taken out of |\ifhmode|, this
% line caused some codeline numbers were typeset with
% |\leftskip|${}=0$.
\everypar=\@xa{%
\@xa\@codetonarrskip\the\gmd@preverypar}%
\let\par\@@par%
\fi}% of \cs{par}.
\gmd@endpe\ignorespaces}}
% As we announced, we play with |\leftskip| inside the verbatim
% group and therefore we wish to restore normal |\leftskip| when back to
% normal text i.e.\ the commentary. But, if normal text starts in the
% same line as the code, then we still wish to indent such a~line.
\def\gmd@endpe{%
\ifprevhmode
\settexcodehangi%\unskip ndent
\leftskip=\CodeIndent
%^^A \typeout{gmd@endpe in prevhmode \on@line}%
\else
\leftskip=\TextIndent
\hangindent=\z@
\everypar=\@xa{%
\@xa\@codetonarrskip\the\gmd@preverypar}%
%^^A \typeout{gmd@endpe in not prevhmode \on@line}%
\fi}
% Now a~special treatment for an inline comment:
\newif\ifilrr
\def\ilrr{%\changes{v0.99n}{2008/08/21}{added}
\if@aftercode
\unless\if@ilgroup\bgroup\@ilgrouptrue\fi% \label{inline.bgroup} If we are
% `aftercode', then we are in an inline comment. Then we open
% a~group to be able to declare e.g.\ \cs{raggedright} for that
% comment only. This group is closed in line \ref{inline.egroup.1}
% or \ref{inline.egroup.2}.
\ilrrtrue
\fi}
\newif\if@ilgroup
\def\gmd@setilrr{\rightskip0ptplus\textwidth}
\def\ilju{% when inline comments are ragged right in general but we
% want just this one to be justified.
\if@aftercode
\unless\if@ilgroup\bgroup\@ilgrouptrue\fi
\ilrrfalse
\fi}
\def\verbcodecorr{%\changes{v0.99n}{2008/08/21}{added}
% a~correction of vertical spaces between a~\env{verbatim} and
% code. We put also a~\cs{par} to allow parindent in the next
% commentary.
\vskip-\lastskip\vskip-4\CodeTopsep\vskip3\CodeTopsep\par}
%
%
% \subdivision{Numbering (or not) of the lines}
%
% Maybe you want codelines to be numbered and maybe you want to
% reset the counter within sections.
\if@uresetlinecount% with |uresetlinecount| option\dots
\@relaxen\gmd@resetlinecount% \dots\ we turn
% resetting the counter by \cs{Doc\+In\+put} off\dots
\newcommand*\resetlinecountwith[1]{%
\newcounter{codelinenum}[#1]}% \dots\ and provide a~new
% declaration of the counter.
\else% With the option turned off\dots
\newcounter{DocInputsCount}%
\newcounter{codelinenum}[DocInputsCount]% \dots\ we declare the
% |\DocInput|s' number counter andthe codeline counter
% to be reset with stepping of it.
% \changes[\c@DocInputsCount]{v0.98c}{06/9/8}{added for fixing
% duplication of \pk{hyperref} labels in the case of a~multiple
% \cs{DocInput}}
\newcommand*\gmd@resetlinecount{\stepcounter{DocInputsCount}}% \dots
% and let the |\DocInput| increment the |\DocInput|s number count
% and thus reset the codeline count. It's for unique naming of the
% \pk{hyperref} labels.
\fi
% Let's define printing the line number as we did in \pk{gmvb}
% package.
\newcommand*\printlinenumber{%
\leavevmode\llap{\rlap{\LineNumFont$\phantom{999}$\llap{\thecodelinenum}}%
\hskip\leftskip}}
\def\LineNumFont{\normalfont\tiny}
\if@linesnotnum\@relaxen\printlinenumber\fi
\newcommand*\hyperlabel@line{%
\if@pageindex% It's good to be able to switch it any time not just
% define it once according to the value of the switch set by the
% option.
\else
\raisebox{2ex}[1ex][\z@]{\gmhypertarget[clnum.%
\HLPrefix\arabic{codelinenum}]{}}%
\fi}
%
% \subdivision{Spacing with \cs{everypar}}
%
% Last but not least, let's define the macro inserting a~vertical space
% between the code and the narration. Its parameter is
% a~\gmhypertarget[relic parameter]{relic} of a~very heavy debug of
% the automatic vspacing mechanism. Let it remain at least until this
% package is 2.0 version.
\newcommand*\gmd@codeskip[1]{%
\@@par\addvspace\CodeTopsep
\@codeskipputgtrue\@nostanzagfalse}
%\label{codeskip2}
% Sometimes we add the |\CodeTopsep| vertical space in
% |\everypar|. When this happens, first we remove the |\parindent|
% empty box, but this doesn't reverse putting |\parskip| to the main
% vertical list. And if |\parskip| is put, |\addvspace| shall see it
% not the `true' last skip. \Describe*{@codeskipput}Therefore we need
% a~Boolean switch to keep the knowledge of putting similar vskip
% before |\parskip|.
% \Define\if@codeskipput
\newgif\if@codeskipput
% A~switch to control |\nostanza|s: \nostanza
\newgif\if@nostanza
% The below is another relic of the heavy debug of the automatic
% vspacing. Let's give it the same removal clause as
% \gmiflink[relic parameter]{above}.
\newcommand*\gmd@nocodeskip[2]{}
% And here is how the two relic macros looked like during the
% debug. As you see, they are disabled by a~false |\if| (look at it
% closely ;-).
\if1 1
\renewcommand*\gmd@codeskip[1]{%
\hbox{\rule{1cm}{3pt} #1!!!}}
\renewcommand*\gmd@nocodeskip[2]{%
\hbox{\rule{1cm}{0.5pt} #1: #2 }}
\fi
% We'll wish to execute |\gmd@codeskip| wherever a~codeline (possibly with
% an inline comment) is followed by a~homogenic comment line or
% reverse. Let us dedicate a~Boolean switch to this then.
% \Define\if@aftercode
\newgif\if@aftercode
% This switch will be set true in the moments when we are able to
% switch from the \TeX\ code into the narration and the below one when
% we are able to switch reversely.
% \Define\if@afternarr
\newgif\if@afternarr
% To insert vertical glue between the \TeX\ code and the narration
% we'll be playing with |\everypar|. More precisely, we'll add a~macro
% that the |\parindent| box shall move and the glue shall put.
\def\@codetonarrskip{%
\if@codeskipput\else
\if@afternarr\gmd@nocodeskip4{iaN}\else
\if@aftercode
% We are at the beginning of |\everypar|, i.e., \TeX\ has just entered
% the hmode and put the |\parindent| box. Let's remove it then.
{\setbox0=\lastbox}%
% Now we can put the vertical space and state we are not `aftercode'.
\gmd@codeskip4%
% \label{codeskip5}
\else\gmd@nocodeskip4{naC}%
\fi
\fi
\fi
\leftskip\TextIndent% this line is a~patch against
% a~bug-or-feature that in certain cases the narration |\leftskip|
% is left equal the code leftskip. (It happens when there're
% subsequent code lines after an inline comment not ended with
% an explicit |\par|.) Before v0.99n it was just after line \ref{codeskip5}.
\@aftercodegfalse\@nostanzagtrue
% \changes{v0.99o}{2008/09/04}{a~bug fix: added \cs{@nostanzagtrue}}
}
% But we play with |\everypar| for other reasons too, and while restoring it,
% we don't want to add the |\@codetonarrskip| macro infinitely many
% times. So let us define a~macro that'll check if |\everypar| begins
% with |\@codetonarrskip| and trim it if so. We'll use this macro with
% proper |\expandafter|ing in order to give it the contents of
% |\everypar|. The work should be done in two steps first of which
% will be checking whether |\everypar| is nonempty (we can't have two
% delimited parameters for a~macro: if we define a~two-parameter
% macro, the first is undelimited so it has to be nonempty; it costed
% me some one hour to understand it).
\long\def\@trimandstore#1\@trimandstore{%
\def\@trimandstore@hash{#1}%
\ifx\@trimandstore@hash\@empty% we check if |#1| is
% nonempty. The \incs{if} \inverb*|\relax#1\relax| trick is not
% recommended here because using it we couldn't avoid expanding |#1|
% if it'd be expandable.
\gmd@preverypar={}%
\else
\afterfi{\@xa\@trimandstore@ne\the\everypar\@trimandstore}%
\fi}
\long\def\@trimandstore@ne#1#2\@trimandstore{%\label{@trimandstore}
\def\trimmed@everypar{#2}%
\ifx\@codetonarrskip#1%
\gmd@preverypar=\@xa{\trimmed@everypar}%
\else
\gmd@preverypar=\@xa{\the\everypar}%
\fi}
% We prefer not to repeat |#1| and |#2| within the |\if|s and we even
% define an auxiliary macro because |\everypar| may contain some
% |\if|s or |\fi|s.
%
%
% \subdivision[Life among queer \acro{EOL}s]{Life among queer \gmhypertarget[afc]{\acro{EOL}s}}
%
% When I~showed this package to my \TeX\ Guru he commended it and
% immediately pointed some disadvantages in the comparison with the
% \docfm\ package.
%
% One of them was an expected difficulty of breaking a~moving argument
% (e.g., of a~sectioning macro) in two lines. To work it around let's
% define a~line-end eater:
\catcode`\^^B=\active% note we re|\catcode| \<char2> globally, for the
% entire document.
\foone{\obeylines}%% \Define*{^^B}
{\def\QueerCharTwo{%
\protected\def^^B##1^^M{%
%^^A\@newlinegtrue\gmd@countnarrline
\ifhmode\unskip\space\ignorespaces\fi}}% It shouldn't be \incs{ } not to
% drive \TeX\ into hmode.
}
\QueerCharTwo
\AtBegInput{\@ifEOLactive{\catcode`\^^B\active}{}\QueerCharTwo}% \label{QCh2}
% We repeat redefinition of \<char2> at begin of the documenting
% input, because \pk{doc.dtx} suggests that some packages (namely
% \pk{inputenc}) may re|\catcode| such unusual characters.
%
% \changes{v0.98d}{06/9/11}{re\cs{catcode}ing and defining
% \cs{char1} and \cs{char2} added atto begin doc input}
%
% \stanza
% As you see the |^^B| active char is defined to gobble everything
% since itself till the end of line and the very end of line. This is
% intended for harmless continuing a~line. The price is affecting the
% line numbering when |countalllines| option is enabled.
%
% I~also liked the \docfm's idea of comment${}^2$ i.e., the
% possibility of marking some text so that it doesn't appear nor in the
% working version neither in the documentation, got by making
% |^^A| (i.e., \<char1>) a~comment char.
%
% However, in this package such a~trick would work another way: here
% the line ends are active, a~comment char would disable them and that
% would cause disasters. So let's do it an |\active| way.
\catcode`\^^A=\active% note we re|\catcode| \<char1> globally, for the
% entire document.
\foone\obeylines{%%\DefIndex\QueerCharOne \Define*{^^A}
\def\QueerCharOne{%
\def^^A{%^^A no need to write \incs{gmd@countnarrline} because
% ^^A \inverb|^^M| will contain it if \inenv{countalllines} is in force.
\bgroup\let\do\@makeother\dospecials\gmd@gobbleuntilM}}%
\def\gmd@gobbleuntilM#1^^M{\egroup\ignorespaces^^M}%
}
\QueerCharOne
\AtBegInput{\@ifEOLactive{\catcode`\^^A\active}\QueerCharOne}% see note
% after line \ref{QCh2}.\ilrr
% As I~suggested in the users' guide, |\StraightEOL| and |\QueerEOL| are
% intended to cooperate in harmony for the user's good. They take care
% not only of redefining the line end but also these little things
% related to it.
%
% One usefulness of |\StraightEOL| is allowing linebreaking of the
% command arguments. Another---making possible executing some code lines
% during the documentation pass.
%
% \changes{v0.98b}{06/09/07}{\cs{QueerM} and \cs{StraightM} renamed
% to \cs{QueerEOL} and \cs{StraightEOL} and other (internal) macros
% ending with \cs[]{M}}
\def\StraightEOL{%\label{StraightEOL}
\catcode`\^^M=5
\catcode`\^^A=14
\catcode`\^^B=14
\def\^^M{\ }}
%^^A \unless\if@countalllines
%^^A \let\StraightEOL\StraightEOLnocount
%^^A \fi
%^^A
\foone\obeylines{%
\def\QueerEOL{%\label{QueerEOL}
\catcode`\^^M=\active%
\let^^M\gmd@textEOL%
\catcode`\^^A=\active%
\catcode`\^^B=\active% I~only re|\catcode| \<char1> and \<char2>
% hoping no one but me is \emph{that} perverse to make them
% |\active| and (re)define. (Let me know if I'm wrong at this point.)
\let\^^M=\gmd@bslashEOL}%
%^^A %
%^^A \if@countalllines%
%^^A \def\StraightEOL{%
%^^A \catcode`\^^M=\active%
%^^A \let^^M\gmd@notsostraightEOL%
%^^A \catcode`\^^A=\active%
%^^A \catcode`\^^B=\active% I~only re|\catcode| \<char1> and \<char2>
%^^A % hoping no one but me is \emph{that} perverse to make them
%^^A % |\active| and (re)define. (Let me know if I'm wrong at this point.)
%^^A \let\^^M=\gmd@bslashEOL}%
%^^A %
%^^A \fi%
}
%^^A % \skiplines
%^^A % They won't work in \verb+|verbatim|+ scopes but the clubs redefine
%^^A % the active lineend to be a~breakable (and visible) space, so
%^^A % breaking a~`clubbed' text in two lines will not cause a~disaster but
%^^A % you get e.g., |...^^B %...|.
%^^A % \endskiplines
% \gmhypertarget[closingspace]{To make} |^^M| behave more like a~`normal'
% lineend I~command it to add a~| |${}\subs{10}$ at first. It works
% but has one unwelcome feature: if the line has nearly |\textwidth|,
% this closing space may cause line breaking and setting a~blank line.
% To fix this I~|\advance| the |\parfillskip|:
\def\gmd@parfixclosingspace{{%
\advance\parfillskip by-\gmd@closingspacewd
\if@aftercode\ifilrr \gmd@setilrr \fi\fi
\par}%
\if@ilgroup\aftergroup\egroup\@ilgroupfalse\fi% we are in the
% verbatim group so we
% close the inline comment group after it if the closing is not yet set.
}
% We'll put it in a~group surrounding |\par| but we need to check if
% this |\par| is executed after narration or after the code,
% i.e., whether the closing space was added or not.
\newskip\gmd@closingspacewd
\newcommand*\gmd@setclosingspacewd{%
\global\gmd@closingspacewd=\fontdimen2\font%
plus\fontdimen3\font minus\fontdimen4\font\relax}
%
% See also line \ref{noclosingspace} to see what we do in the codeline
% case when no closing space is added.
% And one more detail:
\foone\obeylines{%\DefIndex*{\^^M}
\if 1 1%
\protected\def\gmd@bslashEOL{\ \@xa\ignorespaces^^M}%
}% of \cs{foone}. Note we interlace here \incs{if} with a~group.
\else%
\protected\def\gmd@bslashEOL{%
\ifhmode\unskip\fi\ \ignorespaces}
%^^A~\if@countalllines\@newlinegtrue\gmd@cpnarrline\fi%
\fi
% \changes{v0.99c}{2007/03/30}{a~bug fix: redefinition of it left
% solely to \cs{QueerEOL}}
% \changes{v0.99m}{2008/08/09}{also \cs{StraightEOL} with
% \env{countalllines} package option lets \cs{}\hathat{M} to it}
%
% The |\QueerEOL| declaration will |\let| it to |\^^M| to make
% |\^^M| behave properly. If this definition was ommitted, |\^^M|
% would just expand to |\ | and thus not gobble the leading |%| of the
% next line leave alone typesetting the \TeX\ code. I~type |\ | etc.
% instead of just |^^M| which adds a~space itself because I~take
% account of a~possibility of redefining the |\ | \CS by the user, just
% like in normal \TeX.
% We'll need it for restoring queer definitions for \docfm-compatibility.
%
%
% \subdivision{Adjustment of \env{verbatim} and \cs{verb}}
%
% To make \env{verbatim(*)} typeset its contents with the \TeX\ code's
% indentation:
% \Define\@verbatim
\gaddtomacro\@verbatim{\leftskip=\CodeIndent}
% And a~one more little definition to accomodate |\verb| and pals for the
% lines commented out.
\AtBegInput{\long\def\check@percent#1{%
\gmd@cpnarrline% to count the verbatim lines and possibly print
% their numbers. This macro is used only by the verbatim end of line.
\@xa\ifx\code@delim#1\else\afterfi{#1}\fi}}
% We also redefine \pk{gmverb}'s |\AddtoPrivateOthers| that has been
% provided just with \pk{gmdoc}'s need in mind.
\def\AddtoPrivateOthers#1{%
\@xa\def\@xa\doprivateothers\@xa{%
\doprivateothers\do#1}}%
%^^A \foone\obeylines{%
%^^A \AtBegInput{%
%^^A \prependtomacro{% first the stuff then the macro (see \pk{gmutils}).
%^^A \def\obeylines{\def^^M{\par}\catcode`\^^M\active}}%
%^^A \@verbatim}}
% We also redefine an internal |\verb|'s macro |\gm@verb@eol| to put a~proper
% line end if a~line end char is met in a~short verbatim: we have to
% check if we are in `queer' or `straight' \acro{EOL}s area.
\begingroup
\obeylines% \DefIndex\gm@verb@eol
\AtBegInput{\def\gm@verb@eol{\obeylines%
\def^^M{\verb@egroup\@latex@error{%
\@nx\verb ended by end of line}%
\@ifEOLactive{^^M}{\@ehc}}}}%
\endgroup
% \subdivision{Macros for marking of the macros}
%
% A~great inspiration for this part was the \docfm\ package again.
% I~take some macros from it, and some tasks I~solve a~different way,
% e.g., the |\| (or another escapechar) is not active, because anyway
% all the chars of code are scanned one by one. And exclusions from
% indexing are supported not with a~list stored as |\toks| register
% but with separate control sequences for each excluded \CS.
%
% \stanza
%
% The \docfm\ package shows a~very general approach to the indexing
% issue. It assumes using a~special MakeIndex style and doesn't use
% explicit MakeIndex controls but provides specific macros to hide
% them. But here in \pk{gmdoc} we prefer no special style for the
% index.
% \Define\actualchar \Define\quotechar \Define\encapchar
% \Define\levelchar
% \changes[\actualchar]{v0.98e}{06/09/23}{a~bug fix: making the
% \cs[]{@} \protect\catother\space again}
\edef\actualchar{\string @}
\edef\quotechar{\string "}
\edef\encapchar{\xiiclub}
\edef\levelchar{\string !}
% However, for the glossary, i.e., the change history, a~special style
% is required, e.g., \pk{gmglo.ist}, and the above macros are
% redefined by the |\changes| command due to \pk{gmglo.ist} and
% \pk{gglo.ist} settings.
%
% Moreover, if you insist on using a~special MakeIndex style, you may
% redefine the above four macros in the preamble. The |\edef|s that
% process them further are postponed till |\begin{document}|.
%
%
% \Define\code@escape@char
\def\CodeEscapeChar#1{%
\begingroup
\escapechar\m@ne
\xdef\code@escape@char{\string#1}%
\endgroup}
% As you see, to make a~proper use of this macro you should give it
% a~|\|\<one char> \CS as an argument. It's an invariant assertion that
% |\code@escape@char| stores `other' version of the code layer
% escape char.
% \CodeUsgIndex\CodeEscapeChar
\CodeEscapeChar\\
% As mentioned in \docfm, someone may have some chars \catletter ed.
\@ifundefined{MakePrivateLetters}{%\label{MPL}
\def\MakePrivateLetters{\makeatletter\catcode`\*=11 }}{}
% A~tradition seems to exist to write about e.g., `command |\section| and
% command |\section*|' and such an uderstanding also of `macro' is
% noticeable in \docfm. Making the |*| a~letter solves the problem of
% scanning starred commands.
%
% \stanza
% And you may wish some special chars to be \catother.
\def\MakePrivateOthers{\let\do=\@makeother \doprivateothers}
% We use this macro to re|\catcode| the space for marking the
% environments' names and the caret for marking chars such as |^^M|, see
% line \ref{TextUsage}. So let's define the list:
\def\doprivateothers{\do\ \do\^}
% Two chars for the beginning, and also the |\MakeShortVerb| command
% shall this list enlarge with the char(s) declared.
% (There's no need to add the backslash to this list since all the
% relevant commands |\string| their argument whatever it is.)
%
% \changes{v0.98c}{06/9/9}{\cs{} removed from \cs{doprivateothers}
% list and \cs{string} added instead to commands that process the
% `environments' names;}
% Now the main macro indexing a~macro's name. It would be a~verbatim
% :-) copy of the \docfm 's one if I~didn't ommit some lines
% irrelevant with my approach.
\def\scan@macro#1{% we are sure to scan at least one token and
% therefore we define this macro as one-parameter.\par
% Unlike in \docfm , here we have the escape char \catother\ so we
% may just have it printed during main scan char by char, i.e., in the
% lines \ref{printhashone1} and \ref{printhashone2}.\par
% So, we step the checksum counter first,
\step@checksum% \label{checksumUse}(see line \ref{checksum} for
% details),\par
% Then, unlike in \docfm , we do \emph{not} check if the scanning is
% allowed, because here it's always allowed and required.\par
% Of course, I~can imagine horrible perversities, but I~don't think
% they should really be taken into account. Giving the letter |a|
% |\catcode| other than \catletter\ surely would be one of those
% perversities. Therefore I~feel safe to take the character |a| as
% a~benchmark letter.
\ifcat a\@nx#1%
\quote@char#1%
\xdef\macro@iname{\gmd@maybequote#1}% global for symmetry with
% line \ref{x474}.
\xdef\macro@pname{\string#1}%\label{stringing0} we'll print entire
% name of the macro later.\par
% We |\string| it here and in the lines \ref{stringing1} and
% \ref{stringing2} to be sure it is whole \catother\ for easy
% testing for special indexentry formats, see line
% \ref{pnametestDef} etc. Here we are sure the result of |\string|
% is \catother\ since its argument is \catletter.
\afterfi{\@ifnextcat{a}{\gmd@finishifstar#1}{\finish@macroscan}}%
\else% |#1| is not a~letter, so we have just scanned a~one-char
% \CS.\par
% Another reasonable |\catcode|s assumption seems to be that the
% digits are \catother. Then we don't have to
% type (|%|)|\expandafter\@gobble\string\a|. We do the |\uccode|
% trick to be sure that the char we write as the macro's name is
% \catother.
{\uccode`9=`#1%
\uppercase{\xdef\macro@iname{9}}%\label{x474}
}%
\quote@char#1%
\xdef\macro@iname{\gmd@maybequote\macro@iname}%
\xdef\macro@pname{\xiistring#1}%\label{stringing1}
\afterfi \finish@macroscan
\fi}
% The |\xiistring| macro, provided by \pk{gmutils}, is used instead of
% original |\string| because we wish to get | |\catother (`other' space).
%
% \stanza
% Now, let's explain some details, i.e., let's define them. We call
% the following macro having known |#1| to be \catletter.
\def\continue@macroscan#1{%
\quote@char#1%
\xdef\macro@iname{\macro@iname \gmd@maybequote#1}%
\xdef\macro@pname{\macro@pname \string#1}%\label{stringing2} we know
% \inverb*|#1| to be \catletter, so
% we don't need \incs{xiistring}.
\@ifnextcat{a}{\gmd@finishifstar#1}{\finish@macroscan}%
}
% As you may guess, |\@ifnextcat| is defined analogously to |\@ifnextchar| but
% the test it does is |\ifcat| (not |\ifx|). (Note it wouldn't
% work for an active char as the `pattern'.)
%\stanza
% We treat the star specially since in usual \LaTeX\ it should finish
% the scanning of a~\CS name---we want to avoid scanning
% |\command*argum| as one \CS.
\def\gmd@finishifstar#1{%
\if*\@nx#1\afterfi\finish@macroscan% note we protect |#1| against
% expansion. In \pk{gmdoc} verbatim scopes some chars are active
% (e.g. \inverb|\|\,).
\else\afterfi\continue@macroscan
\fi}
% If someone \emph{really} uses |*| as a~letter please let me know.
\def\quote@char#1{{\uccode`9=`#1% at first I~took digit 1 for this
% |\uccode|ing but then |#1| meant |#|\<\#1> in |\uppercase|'s
% argument, of course.
\uppercase{% \DefIndex\gmd@maybequote
\gmd@ifinmeaning 9\of \indexcontrols
{\glet\gmd@maybequote\quotechar}%
{\g@emptify\gmd@maybequote}%
}%
}}
% And now let's take care of the MakeIndex control characters. We'll
% define a~list of them to check whether we should quote a~char or
% not. But we'll do it at |\begin{document}| to allow the user to use
% some special MakeIndex style and in such a~case to redefine the four
% MakeIndex controls' macros. We enrich this list with the backslash
% because sometimes MakeIndex didn't like it unquoted.
% \Define\indexcontrols
% \changes{v0.98d}{06/9/15}{relativized to the index control macros:
% previously the controls were given explicitly as the standard ones.}
\AtBeginDocument{\xdef\indexcontrols{%
\bslash\levelchar\encapchar\actualchar\quotechar}}
\long\def \gmd@ifinmeaning#1\of#2#3#4{% explained in the text
% paragraph below. \changes{v0.99g}{2007/11/06}{made more elegant:
% \cs{if} changed to \cs{ifx} made four parameters and not expanding
% to an open \cs{iftrue/false}. Also renamed from \cs{@ifismember}}
\long\def\gmd@in@@##1#1##2\gmd@in@@{%
\ifx^^A##2^^A\afterfi{#4}%
\else\afterfi{#3}%
\fi}%
\@xa\gmd@in@@#2#1\gmd@in@@}%
% This macro is used for catching chars that are
% MakeIndex's controls. How does it work?
%
% |\quote@char| sort of re|\catcode|s its argument through the
% |\uccode| trick: assigns the argument as the uppercase code of the
% digit 9 and does further work in the |\uppercase|'s scope so the
% digit 9 (a~benchmark `other') is substituted by |#1| but the
% |\catcode| remains so \cs{gmd\-@\-if\-in\-mean\-ing} gets |\quote@char|'s |#1|
% `other'ed as the first argument.
%
% The meaning of the |\gmd@ifinmeaning| parameters is as follows:
%\begin{enumargs}^^B
%\item the token(s) whose presence we check,
%\item the macro in whose meaning we search |#1| (the first token of
% this argument is expanded one level with |\expandafter|),
%\item the `if found' stuff,
%\item the `if not found' stuff.
%\end{enumargs}
%
% In |\quote@char| the second argument for |\gmd@ifinmeaning| is
% |\indexcontrols| defined as the (expanded and `other') sequence of
% the MakeIndex controls. |\gmd@ifinmeaning| defines its inner macro
% |\gmd@in@@| to take two parameters separated by the first and the
% second |\gmd@ifinmeaning|'s parameter, which are here the char
% investigated by |\quote@char| and the |\indexcontrols| list. The
% inner macro's parameter string is delimited by the macro itself, why
% not. |\gmd@in@@| is put before a~string consisting of
% |\gmd@ifinmeaning|'s second and first parameters (in such a~reversed
% order) and |\gmd@in@@| itself. In such a~sequence it looks for
% something fitting its parameter pattern. |\gmd@in@@| is sure to find
% the parameters delimiter (|\gmd@in@@| itself) and the separator,
% |\ifismember|'s |#1| i.e., the investigated char, because they are
% just there. But the investigated char may be found not near the end,
% where we put it, but among the MakeIndex controls' list. Then the
% rest of this list and |\ifismember|'s |#1| put by us become the
% secong argument of |\gmd@in@@|. What |\gmd@in@@| does with its
% arguments, is just a~check whether the second one is empty. This may
% happen \emph{iff} the investigated char hasn't been found among the
% MakeIndex controls' list and then |\gmd@in@@| shall expand to
% |\iffalse|, otherwise it'll expand to |\iftrue|. (The |\after...|
% macros are employed not to (mis)match just got |\if...| with the
% test's |\fi|.) ``(Deep breath.) You got that?'' If not, try
% \docfm's explanation of |\ifnot@excluded|, pp.\,36--37 of the v2.1b
% dated 2004/02/09 documentation, where a~similar construction is
% attributed to Michael Spivak.
%
% Since version 0.99g |\gmd@ifinmeaning| is used also in testing
% whether a~detector is already present in the carrier in the
% mechanism of automatic detection of definitions (line \ref{550}).
\newif\ifgmd@glosscs% we use this switch to keep the information
% whether a~history entry is a~\CS or not.
\newcommand*\finish@macroscan{%\label{506}\par
% First we check if the current \CS is not just being defined. The
% switch may be set true in line \ref{519}
\ifgmd@adef@cshook% if so, we throw it into marginpar and index as
% a~def entry\dots
\@ifundefined{gmd/iexcl/\macro@pname}{% \dots\ if it's not excluded
% from indexing.
\@xa\Code@MarginizeMacro\@xa{\macro@pname}%
\@xa\@defentryze\@xa{\macro@pname}{1}}{}%% here we declare the kind of
% index entry and define |\last@defmark| used by \cs{changes}
\global\gmd@adef@cshookfalse% we falsify the hook that was set
% true just for this \CS.
\fi
% We have the \CS's name for indexing in |\macro@iname| and
% for print in |\macro@pname|. So we index it. We do it a~bit
% countercrank way because we wish to use more general indexing
% macro.
\if\verbatimchar\macro@pname% \label{3039}it's important that |\verbatimchar|
% comes before the macro's name: when it was reverse, the |\tt| \CS
% turned this test true and left the |\verbatimchar| what resulted
% with `|\+tt|' typeset. Note that this test should turn true iff
% the scanned macro name shows to be the default
% |\verb|'s delimiter. In such a~case we give
% |\verb| another delimiter, namely |$|: ^^A$
\def\im@firstpar{[$%^^A$
]}%
\else\def\im@firstpar{}\fi
\@xa \index@macro\im@firstpar\macro@iname\macro@pname
% \label{3049}
\maybe@marginpar\macro@pname
\if\xiispace\macro@pname\relax\gmd@texcodespace
\else\macro@pname
\fi
% \changes[\finish@macroscan]{v0.99n}{2008/09/30}{the case of
% \cs{\vs} taken care of}
\let\next\gmd@charbychar
\gmd@detectors% \label{519} for automatic detection of
% definitions. Defined and
% explained in the next section. It redefines
% |\next| if detects a~definition command and thus
% sets the switch of line \ref{506} true.
\next
% \label{next 3690}
}
% Now, the~macro that checks whether the just scanned macro should be
% put into a~marginpar: it checks the meaning of a~very special \CS:
% whose name consists of |gmd/2marpar/| and of the examined macro's
% name.
\def\maybe@marginpar#1{%
%^^A\ifx#1\mname@tomarginpar
\@ifundefined{gmd/2marpar/#1}{}{%
\@xa\Text@Marginize\@xa{\bslash#1}% |\expandafter|s
% \possfil because the |\Text@Marginize| command applies |\string| to its
% argument. \incs{macro@pname}, which will be the only possible
% argument to
% \incs{may\+be\+@mar\+g\+in\+par},
% contains the macro's name
% without the escapechar so we added it here.
% ^^A\gdef\mname@tomarginpar{}
\@xa\g@relaxen\csname gmd/2marpar/#1\endcsname% we reset the switch.
}}
% Since version 0.99g we introduce automatic detection of definitions,
% it will be implemented in the next section. The details of indexing
% \CSs are implemented in the section after it.
%
% \subdivision{Automatic detection of definitions}
%
% To begin with, let's introduce a~general declaration of a~defining
% command. \cs{Decl\-are\-Def\-in\-ing} comes in two flavours: `saut\eacute',
% and with star. The `saut\eacute' version without an optional
% argument declares a~defining command of the kind of |\def| and
% |\newcommand|: whether wrapped in braces or not, its main argument
% is a~\CS. The star version without the optional argument declares
% a~defining command of the kind of |\newenvironment| and
% |\DeclareOption|: whose main mandatory argument is text. Both
% versions provide an optional argument in which you can set the keys.
% Probably the most important key is |star|. It determines whether the
% starred version of a~defining command should be taken into account.
% For example, |\newcommand| should be declared with |[star=true]|
% while |\def| with |[star=false]|. You can also write just |[star]|
% instead of |[star=true]|. It's the default if the |star| key is
% omitted.
%
% Another key is |type|. Its possible values are the (backslashless)
% names of the defining commands, see below.
%
% We provide now more keys for the \pk{xkeyval}ish definitions:
% |KVpref| (the key prefix) and |KVfam| (the key family). If not set by
% the user, they are assigned the default values as in \pk{xkeyval}:
% |KVpref| letters |KV| and |KVfam| the input file name. The latter
% assignment is done only for the |\DeclareOptionX| defining command
% because in other \pk{xkeyval} definitions (|\define@(...)key|) the
% family is mandatory.
%
%\stanza
% Let's make a~version of |\@ifstar| that would work with
% |*|\catletter. It's analogous to |\@ifstar|.
\foone{\catcode`\*=11 }
{\def\@ifstarl#1{\@ifnextchar *{\@firstoftwo{#1}}}}
%
% \subsubdivision{\cs{DeclareDefining} and the detectors}
%
% Note that the main argument of the next declaration should be a~\CS
% \emph{without star}, unless you wish to declare only the starred
% version of a~command. The effect of this command is always global.
\outer\def\DeclareDefining{\begingroup
\MakePrivateLetters
\@ifstarl
{\gdef\gmd@adef@defaulttype{text}\Declare@Dfng}%
{\gdef\gmd@adef@defaulttype{cs}\Declare@Dfng}%
}
% The keys except |star| depend of |\gmd@adef@currdef|, therefore we
% set them having known both arguments
\newcommand*\Declare@Dfng[2][]{%
\endgroup
\Declare@Dfng@inner{#1}{#2}%
\ifgmd@adef@star% this swith may be set false in first
% \incs{Declare@Dfng@inner} (it's the |star| key).
\Declare@Dfng@inner{#1}{#2*}% The catcode of |*| doesn't matter since
% it's in
% \incs{csname\+…\+\bslash end\+cs\+na\+me}
% everywhere.
\fi}
% \Define\gmd@adef@currdef
\def\Declare@Dfng@inner#1#2{%
\edef\gmd@resa{%
\@nx\setkeys[gmd]{adef}{type=\gmd@adef@defaulttype}}%
\gmd@resa
{\escapechar\m@ne
\xdef\gmd@adef@currdef{\string#2}%
% ^^A~\typeout{@@@ gmd@adef@currdef:::\gmd@adef@currdef::::}%
}%
\gmd@adef@setkeysdefault
\setkeys[gmd]{adef}{#1}%
\@xa\gmd@ifinmeaning
\csname gmd@detect@\gmd@adef@currdef\endcsname
% \label{550}
\of\gmd@detectors{}{%
\@xa\gaddtomacro\@xa\gmd@detectors\@xa{%
\csname gmd@detect@\gmd@adef@currdef\endcsname}}% we add a~\CS\\
% |%| |\gmd@detect@|\<def name> (a~\textbf{detector}) to the
% meaning of the \textbf{detectors' carrier}. And we define it to
% detect the \inverb|#2| command.
\@xa\xdef\csname gmd@detectname@\gmd@adef@currdef\endcsname{%
\gmd@adef@currdef}%
\edef\gmu@tempa{% this |\edef| is to expand |\gmd@adef@TYPE|.
\global\@nx\@namedef{gmd@detect@\gmd@adef@currdef}{%
\@nx\ifx
\@xa\@nx\csname gmd@detectname@\gmd@adef@currdef\endcsname
\@nx\macro@pname
\@nx\n@melet{next}{gmd@adef@\gmd@adef@TYPE}%
\@nx\n@melet{gmd@adef@currdef}{gmd@detectname@\gmd@adef@currdef}%
\@nx\fi}}%
\gmu@tempa
\SMglobal\StoreMacro*{gmd@detect@\gmd@adef@currdef}% we store the \CS to
% allow its temporary discarding later.
}
\def\gmd@adef@setkeysdefault{%
\setkeys[gmd]{adef}{star,prefix,KVpref}}
% Note we don't set |KVfam|. We do not so because for
% |\define@key|-likes family is a~mandatory argument and for
% |\DeclareOptionX| the default family is set to the input file name
% in line \ref{defDOXfam}.
\define@boolkey[gmd]{adef}{star}[true]{}
% The |prefix@|\<command> keyvalue will be used to create additional
% index entry for detected definiendum (a~\textbf{definiendum} is the
% thing defined, e.g. in |\newenvironment{foo}| the env.\ \env{foo}).
% For instance, |\newcounter| is declared with |[prefix=\bslash c@]|
% in line \ref{newcounter} and therefore |\newcounter{foo}| occurring
% in the code will index both |foo| and |\c@foo| (as definition
% entries). \UnDef
\define@key[gmd]{adef}{prefix}[]{%
\edef\gmd@resa{%
\def\@xa\@nx\csname gmd@adef@prefix@\gmd@adef@currdef \endcsname{%
#1}}%
\gmd@resa}
\def\gmd@KVprefdefault{KV}% in a~separate macro because we'll need
% it in \cs{ifx}.
% A~macro |\gmd@adef@KVprefixset@|\<command> if defined, will falsify
% an \cs{ifnum} test that will decide whether create additional index
% entry together with the tests for |prefix|\<command> and
% \UnDef
\define@key[gmd]{adef}{KVpref}[\gmd@KVprefdefault]{%
\edef\gmd@resa{#1}%
\ifx\gmd@resa\gmd@KVprefdefault
\else
\@namedef{gmd@adef@KVprefixset@\gmd@adef@currdef}{1}%
\gmd@adef@setKV% whenever the |KVpref|fix is set (not default), the
% declared command is assumed to be \pk{keyval}ish.
\fi
\edef\gmd@resa{#1}% because |\gmd@adef@setKV| redefined it.
\edef\gmd@resa{%
\def\@xa\@nx\csname gmd@adef@KVpref@\gmd@adef@currdef\endcsname{%
\ifx\gmd@resa\empty
\else#1@\fi}}% as in \pk{xkeyval}, if the \acro{KV} prefix is not
% empty, we add \inverb|@| to it.
\gmd@resa}
% Analogously to |KVpref|, |KVfam| declared in |\DeclareDefining|
% will override the family scanned from the code and, in
% |\DeclareOptionX| case, the default family which is the input file
% name (only for the command being declared).
\define@key[gmd]{adef}{KVfam}[]{%
\edef\gmd@resa{#1}%
\@namedef{gmd@adef@KVfamset@\gmd@adef@currdef}{1}%
\edef\gmd@resa{%
\def\@xa\@nx\csname gmd@adef@KVfam@\gmd@adef@currdef\endcsname{%
\ifx\gmd@resa\empty
\else#1@\fi}}%
\gmd@resa
\gmd@adef@setKV}% whenever the |KVfam|ily is set, the declared command is
% assumed to be \pk{keyval}ish.
\define@choicekey[gmd]{adef}{type}
[\gmd@adef@typevals\gmd@adef@typenr]
{% the list of possible types of defining commands
def,
newcommand,
cs,% equivalent to the two above, covers all the cases of defining
% a~\CS, including the \PlainTeX\ \inverb|\new...| and
% \LaTeX\ |\newlength|.
newenvironment,
text,% equivalent to the one above, covers all the commands defining
% its first mandatory argument that should be text,
% \inverb|\DeclareOption| e.g.
define@key,% special case of more arguments important; covers the
% \pk{xkeyval} defining commands.
dk,% a~shorthand for the one above.
DeclareOptionX,% another case of special arguments configuration,
% covers the \pk{xkeyval} homonym.
dox,% a~shorthand for the one above.
kvo% one of option defining commands of the \pk{kvoptions} package
% by Heiko Oberdiek (a~package available o~\acro{CTAN} in the
% \pk{oberdiek} bundle).
}
{% In fact we collapse all the types just to four so far:
\ifcase\gmd@adef@typenr% if |def|
\gmd@adef@settype{cs}{0}%
\or% when |newcommand|
\gmd@adef@settype{cs}{0}%
\or% when |cs|
\gmd@adef@settype{cs}{0}%
\or% when |newenvironment|
\gmd@adef@settype{text}{0}%
\or% when |text|
\gmd@adef@settype{text}{0}%
\or% when |define@key|
\gmd@adef@settype{dk}{1}%
\or% when |dk|
\gmd@adef@settype{dk}{1}%
\or% when |DeclareOptionX|
\gmd@adef@settype{dox}{1}%
\or% when |dox|
\gmd@adef@settype{dox}{1}%
\or% when |kvo|
\gmd@adef@settype{text}{1}%% The \pk{kvoptions} option
%% definitions take first mandatory
% argument as the option name and they define a~\pk{keyval} key
% whose macro's name begins with the prefix/family, either default or
% explicitly declared. The \pk{kvoptions} prefix/family is
% supported in \pk{gmdoc} with \inverb|[KVpref=, KVfam=|\<family>|]|.
\fi}
\def\gmd@adef@settype#1#2{%
\def\gmd@adef@TYPE{#1}%
\ifnum1=#2 % now we define (or not) a~quasi-switch that fires for
% the \pk{keyval}ish definition commands.
\gmd@adef@setKV
\fi}
\def\gmd@adef@setKV{%
\edef\gmd@resa{%
\def\@xa\@nx\csname gmd@adef@KV@\gmd@adef@currdef\endcsname{1}%
}%
\gmd@resa}
% We initialize the carrier of detectors:
\emptify\gmd@detectors
% The definiendum of a~command of the |cs| type is the next control
% sequence. Therefore we only need a~self-relaxing hook in
% |\finish@macroscan|.
\newif\ifgmd@adef@cshook
\def\gmd@adef@cs{\global\gmd@adef@cshooktrue\gmd@charbychar}
% For other kinds of definitions we'll employ active chars of their
% arguments' opening braces, brackets and seargants. In \pk{gmdoc}
% code layer scopes the left brace is active so we only add a~hook to
% its meaning (see line \ref{gm@lbracehook} in \pk{gmverb}) and \ref and here
% we switch it according to the type of detected definition.
\def\gmd@adef@text{\gdef\gmd@lbracecase{1}\gmd@charbychar}
\foone{%
\catcode`\[\active
% ^^A>\]
\catcode`\<\active}
{%\par
% The detector of \pk{xkeyval} |\define@(...)key|:
\def\gmd@adef@dk{%
\let[\gmd@adef@scanKVpref
\catcode`\[\active
% ^^A\]]
\gdef\gmd@lbracecase{2}%
\gmd@adef@dfKVpref\gmd@KVprefdefault% We set the default value of
% the \pk{xkeyval} prefix. Each time again because an assignment
% in \inverb|\gmd@adef@dfKVpref| is global.\ilrr
\gmd@adef@checklbracket}
% The detector of \pk{xkeyval} |\DeclareOptionX|:
\def\gmd@adef@dox{%
\let[\gmd@adef@scanKVpref
\let<\gmd@adef@scanDOXfam
\catcode`[\active
% ^^A]]
\catcode`<\active
\gdef\gmd@lbracecase{1}%
\gmd@adef@dfKVpref\gmd@KVprefdefault% We set the default values of
% the \pk{xkeyval} prefix\dots
\edef\gmd@adef@fam{\gmd@inputname}% \dots\ and family.
\gmd@adef@dofam
% \label{defDOXfam}
\gmd@adef@checkDOXopts}%
}
% The case when the right bracket is next to us is special because it
% is already touched by |\futurelet| (of \CSs scanning macro's
% \cs{@ifnextcat}), therefore we need a~`future' test.
\def\gmd@adef@checklbracket{%
\@ifnextchar[%^^A]
\gmd@adef@scanKVpref\gmd@charbychar}% note that
% the prefix scanning macro gobbles its first argument (undelimited)
% which in this case is |[|.\ilrr
%^^A]
% After a~|\DeclareOptionX|-like defining command not only the prefix
% in square brackets may occur but also the family in
% seargants. Therefore we have to test presence of both of them.
\def\gmd@adef@checkDOXopts{%
\@ifnextchar[\gmd@adef@scanKVpref%^^A]
{\@ifnextchar<\gmd@adef@scanDOXfam\gmd@charbychar}}
%^^A[
\def\gmd@adef@scanKVpref#1#2]{%
\gmd@adef@dfKVpref{#2}%
[#2]\gmd@charbychar}
\def\gmd@adef@dfKVpref#1{%
\ifnum1=0\csname gmd@adef@KVprefixset@\gmd@adef@currdef\endcsname
\relax
\else
\edef\gmu@resa{%
\gdef\@xa\@nx
\csname gmd@adef@KVpref@\gmd@adef@currdef\endcsname{%
\ifx\relax#1\relax
\else#1@%
\fi}}%
\gmu@resa
\fi}
\def\gmd@adef@scanDOXfam{%
\ifnum12=\catcode`\>\relax
\let\next\gmd@adef@scanfamoth
\else
\ifnum13=\catcode`\>\relax
\let\next\gmd@adef@scanfamact
\else
\PackageError{gmdoc}{> neither `other' nor `active'! Make it
`other' with \bslash AddtoPrivateOthers\bslash\>.}%
\fi
\fi
\next}
\def\gmd@adef@scanfamoth#1>{%
\edef\gmd@adef@fam{\@gobble#1}% there is always
% \cs{gmd@charbychar} first.
\gmd@adef@dofam
<\gmd@adef@fam>%
\gmd@charbychar}
\foone{\catcode`\>\active}
{\def\gmd@adef@scanfamact#1>{%
\edef\gmd@adef@fam{\@gobble#1}% there is always
% \cs{gmd@charbychar} first.
\gmd@adef@dofam
<\gmd@adef@fam>%
\gmd@charbychar}%
}
% The hook of the left brace consists of \cs{ifcase} that logically
% consists of three subcases:
% \begin{itemize}^^B
% \item [0]---the default: do nothing in particular;
% \item [1]---the detected defining command has one mandatory
% argument (is of the |text| type, including \pk{kvoptions} option definition);
% \item [2--3]---we are after detection of a~|\define@key|-like command
% so we have to scan \emph{two} mandatory arguments (case 2 is for
% the family, case 3 for the key name).
% \end{itemize}
%
%
\def\gm@lbracehook{%
\ifcase\gmd@lbracecase\relax
\or% when 1
\afterfi{%
\gdef\gmd@lbracecase{0}%
\gmd@adef@scanname}%
\or% when 2---the first mandatory argument of two (|\define@(...)key|)
\afterfi{%
\gdef\gmd@lbracecase{3}%
\gmd@adef@scanDKfam}%
\or% when 3---the second mandatory argument of two (the key name).
\afterfi{%
\gdef\gmd@lbracecase{0}%
\gmd@adef@scanname}%
\fi}
\def\gmd@lbracecase{0}% we initialize the hook caser.
% And we define the inner left brace macros:
\foone{\catcode`\[1 \catcode`\]2 \catcode`\}12 }
[% Note that till line \ref{738} the square brackets are grouping
% and the right brace is `other'. ^^A{
%\par Define the macro that reads and processes the |\define@key|
%family argument. It has the parameter delimited with `other' right
%brace. An active left brace that has launched this macro had been
%passed through iterating |\gmd@charbychar| that now stands next right
%to us.
\def\gmd@adef@scanDKfam#1}[%^^A{
\edef\gmd@adef@fam[\@gobble#1]% there is always
% \cs{gmd@charbychar} first.
\gmd@adef@dofam
\gmd@adef@fam}%
\gmd@charbychar]
% ^^A{
\def\gmd@adef@scanname#1}[%^^A{
\@makeother\[%^^A\]
\@makeother\<%
% The scanned name begins with |\gmd@charbychar|, we have to be
% careful.
\gmd@adef@deftext[#1]%
\@gobble#1}%
\gmd@charbychar]
]
\def\gmd@adef@dofam{%
\ifnum1=0\csname gmd@adef@KVfamset@\gmd@adef@currdef\endcsname
\relax% a~family declared with |\DeclareDefining| overrides the
% one currently scanned.
\else
\edef\gmu@resa{%
\gdef\@xa\@nx
\csname gmd@adef@KVfam@\gmd@adef@currdef\endcsname
{\ifx\gmd@adef@fam\empty
\else\gmd@adef@fam @%
\fi}}%
\gmu@resa
\fi}
\def\gmd@adef@deftext#1{%
\edef\macro@pname{\@gobble#1}% we gobble |\gmd@charbychar|, cf. above.
\@xa\Text@Marginize\@xa{\macro@pname}%
\gmd@adef@indextext
\edef\gmd@adef@altindex{%
\csname gmd@adef@prefix@\gmd@adef@currdef \endcsname}%
%\hskip-\parindent and we add the \pk{xkeyval} header if we are in \pk{xkeyval}
% definition.
\ifnum1=0\csname gmd@adef@KV@\gmd@adef@currdef \endcsname\relax% The\\ \CS
% \inverb|\gmd@adef@KV@|\<def. command> is defined |{1}| (so \cs{ifnum}
% gets |1=01\relax|---\hskip0sptrue) iff \<def. command> is a~\pk{keyval}
% definition. In that case we check for the \inverb|KVpref|ix and
% \inverb|KVfam|ily. (Otherwise |\gmd@adef@KV@|\<def. command> is undefined
% so \cs{ifnum} gets |1=0\relax|---false.)\ilrr
\edef\gmd@adef@altindex{%
\gmd@adef@altindex
\csname gmd@adef@KVpref@\gmd@adef@currdef \endcsname}%
\edef\gmd@adef@altindex{%
\gmd@adef@altindex
\csname gmd@adef@KVfam@\gmd@adef@currdef \endcsname}%
\fi
\ifx\gmd@adef@altindex\empty
\else% we make another index entry of the definiendum with prefix/KVheader.
\edef\macro@pname{\gmd@adef@altindex\macro@pname}%
\gmd@adef@indextext
\fi}
\def\gmd@adef@indextext{%
\@xa\@defentryze\@xa{\macro@pname}{0}% declare the definiendum has to
% have a~definition entry and in the changes history should appear
% without backslash.
\gmd@doindexingtext% redefine |\do| to an indexing macro.
% \label{gmd@doindexingtext 2nd use}
\@xa\do\@xa{\macro@pname}}
% \stanza
% So we have implemented automatic detection of definitions. Let's now
% introduce some.
%
% \subsubdivision{Default defining commands}
% Some commands are easy to declare as defining:
% \HideAllDefining
% \DeclareDefining\DeclareDefining
%^^A\DeclareDefining[type=dox, star=false]\def
\DeclareDefining[star=false]\def
\DeclareDefining[star=false]\pdef% it's a~\pk{gmutils}' shorthand for \inverb|\protected\def|.
\DeclareDefining[star=false]\provide% a~\pk{gmutils}' conditional \incs{def}.
\DeclareDefining[star=false]\pprovide% a~\pk{gmutils}' conditional \incs{pdef}.
%^^A\show\gmd@detectors
% \ResumeAllDefining But |\def| definitely \emph{not always} defines
% an important macro. Sometimes it's just a~scratch assignment.
% Therefore we define the next declaration. It turns the next
% occurence of |\def| off (only the next one).
\def\UnDef{{% \changes{v0.99n}{2008/08/30}{a~bug fixed:
% \cs{gmd@charbychar} appended to \cs{next}---without it
% a~subsequent inline comment was typeset verbatim}
% \UnDef
\gmd@adef@selfrestore\def
}}
\StoreMacro\UnDef% because the `hiding' commands relax it.
\def\HideDef{% \changes{v0.99n}{2008/08/30}{added the starred version
% that calls \cs{UnDef}}
\@ifstar\UnDef{\HideDefining\def\relaxen\UnDef}}
\def\ResumeDef{\ResumeDefining\def\RestoreMacro\UnDef}
% Note that I~\emph{don't} declare |\gdef|, |\edef| neither
% |\xdef|. In my opinion their use as `real' definition is very rare
% and then you may use |\Define| implemented later.
% \HideAllDefining \DeclareDefining\DeclareDefining
\DeclareDefining[star=false]\newcount
\DeclareDefining[star=false]\newdimen
\DeclareDefining[star=false]\newskip
\DeclareDefining[star=false]\newif
\DeclareDefining[star=false]\newtoks
\DeclareDefining[star=false]\newbox
\DeclareDefining[star=false]\newread
\DeclareDefining[star=false]\newwrite
\DeclareDefining[star=false]\newlength
\DeclareDefining[star=false]\DeclareDocumentCommand
% \changes{v0.99p}{2008/10/04}{added}
\DeclareDefining\newcommand
\DeclareDefining\renewcommand
\DeclareDefining\providecommand
\DeclareDefining\DeclareRobustCommand
\DeclareDefining\DeclareTextCommand
\DeclareDefining\DeclareTextCommandDefault
\DeclareDefining*\newenvironment
\DeclareDefining*\renewenvironment
\DeclareDefining*\DeclareOption
%|%\DeclareDefining*\@namedef|
% \HideDefining\DeclareDefining
% \DeclareDefining\bslash
\DeclareDefining*[prefix=\bslash c@]\newcounter% \label{newcounter}
% this prefix provides indexing also |\c@|\<counter>.
\DeclareDefining[type=dk, prefix=\bslash]\define@key
\DeclareDefining[type=dk, prefix=\bslash if]\define@boolkey% the
% alternate index entry will be
% \cs{if}\<KVpref>|@|\<KVfam>|@|\<key name>
\DeclareDefining[type=dk, prefix=\bslash]\define@choicekey
\DeclareDefining[type=dox, prefix=\bslash]\DeclareOptionX% the
% alternate index entry will be \cs{}\<KVpref>|@|\<KVfam>|@|\<option name>.
%
% For |\DeclareOptionX| the default KVfamily is the input file
% name. If the source file name differs from the name of the goal
% file (you \TeX\ a~\file{.dtx} not \file{.sty} e.g.), there is the
% next declaration. It takes one optional and one mandatory
% argument. The optional is the |KVpref|, the mandatory the |KVfam|.
% \Define\DeclareDOXHead
\newcommand*\DeclareDOXHead[2][\gmd@KVprefdefault]{%
\csname DeclareDefining\endcsname
[type=dox, prefix=\bslash, KVpref=#1, KVfam=#2]% \HideDefining\DeclareOptionX
\DeclareOptionX
}
% An example:
%\skiplines
\iffalse
%\endskiplines
\DeclareOptionX[Berg]<Lulu>{EvelynLear}{}
% Check in the index for
% \inverb|EvelynLear| and \inverb|\Berg@Lulu@EvelynLear|.
% Now we set in the comment layer
% |\DeclareDOXHead[Webern]{Lieder}|\DeclareDOXHead[Webern]{Lieder} and
\DeclareOptionX<AntonW>{ChneOelze}
% The latter example shows also overriding the option header by
% declaring the default. By the way, both the example options are
% not declared in the code actually.
%\skiplines
\fi
%\endskiplines
%\stanza
% Now the Heiko Oberdiek's \pk{kvoptions} package option definitions:
\DeclareDefining[type=kvo, prefix=\bslash, KVpref=]\DeclareStringOption
\DeclareDefining[type=kvo, prefix=\bslash, KVpref=]\DeclareBoolOption
\DeclareDefining[type=kvo, prefix=\bslash, KVpref=]\DeclareComplementaryOption
\DeclareDefining[type=kvo, prefix=\bslash, KVpref=]\DeclareVoidOption
% The \pk{kvoptions} option definitions allow setting the default
% family/prefix for all definitions forth so let's provide analogon:
\def\DeclareKVOFam#1{%
\def\do##1{%
\csname DeclareDefining\endcsname
[type=kvo, prefix=\bslash, KVpref=, KVfam=#1]##1}%
\do\DeclareStringOption
\do\DeclareBoolOption
\do\DeclareComplementaryOption
\do\DeclareVoidOption
}
% \ResumeAllDefining^^A~ it cancels |\DD\DD| and |\DD\bslash| effect.
% As a~nice exercise I~recommend to think why this list of declarations
% had to be preceded (in the comment layer) with |\HideAllDefining|
% and for which declarations of the above
% |\DeclareDefining\DeclareDefining| did not work. (The answers are
% commented out in the source file.)
%^^A~ I~had to precede the list of |\DeclareDefining|s with
%^^A~ |\HideAllDefining| in the commentary because the detectors of
%^^A~ some commands expect certain configurations of arguments and of
%^^A~ course such arguments do \emph{not} occur when such a~defining
%^^A~ command is being declared. That caused errors. Another reason is
%^^A~ that if not hidden, the detectors fired at the declared definers
%^^A~ of \CSs and marked next |\DeclareDefining| as being defined.
%^^A
%^^A~ |\DD\DD| did not work for those three \pk{xkeyval} definers
%^^A~ declared with |\bslash| in the optional argument. To be precise,
%^^A~ it \emph{did work} perfectly: marked |\bslash| as being defined.
%
% One remark more: if you define (in the code) a~new defining command
% (I~did: a~shorthand for |\DeclareOptionX[gmcc]<>|), declare it as
% defining (in the commentary) \emph{after} it is defined. Otherwise
% its first occurence shall fire the detector and mark next \CS or
% worse, shall make the detector expect some arguments that it won't
% find.
% \subsubdivision{Suspending (`hiding') and resuming detection}
%
% Sometimes we want to suspend automatic detection of definitions. For
% |\def| we defined suspending and resuming declarations in the
% previous section. Now let's take care of detection more generally.
%
% \stanza
% The next command has no arguments and suspends entire detection of
% definitions.
\def\HideAllDefining{%
\ifnum0=0\csname gmd@adef@allstored\endcsname
\SMglobal\StoreMacro\gmd@detectors
\global\@namedef{gmd@adef@allstored}{1}%
\fi
\global\emptify\gmd@detectors}% we make the carrier |\empty| not |\relax| to
% be able to declare new defining command in
% the scope of |\HideAll...|
% The |\ResumeAllDefining| command takes no arguments and restores the
% meaning of the detectors' carrier stored with |\HideAllDefining|
\def\ResumeAllDefining{%
\ifnum1=0\csname gmd@adef@allstored\endcsname\relax
\SMglobal\RestoreMacro\gmd@detectors
\SMglobal\RestoreMacro\UnDef
\global\@namedef{gmd@adef@allstored}{0}%
\fi}
% Note that |\ResumeAllDefining| discards the effect of any
% |\DeclareDefining| that could have occured between
% |\HideAllDefining| and itself.
%
% \stanza
%
% The |\HideDefining| command takes one argument which should be
% a~defining command (always without star). |\HideDefining| suspends
% detection of this command (also of its starred version) until
% |\ResumeDefining| of the same command or |\ResumeAllDefining|.
\def\HideDefining{\begingroup
% \changes{v0.99n}{2008/08/30}{Added the starred version that hides
% the defining command only once}
\MakePrivateLetters
\@ifstarl\Hide@DfngOnce\Hide@Dfng}
\def\Hide@Dfng#1{%%\UnDef
\escapechar\m@ne
\gn@melet{gmd@detect@\string#1}{relax}%
\gn@melet{gmd@detect@\string#1*}{relax}%
\ifx\def#1\global\relaxen\UnDef\fi
\endgroup}
\def\Hide@DfngOnce#1{%%\UnDef
\gmd@adef@selfrestore#1%
\endgroup}
\def\gmd@adef@selfrestore#1{%
\escapechar\m@ne
\@ifundefined{gmd@detect@\string#1}{%
\SMglobal\@xa\StoreMacro
\csname gmd@detect@\string#1\endcsname}{}%
% ^^A\typeout{:::::::::::::::gmd@detect@\string#1:::::::::::::}%
\global\@nameedef{gmd@detect@\string#1}{%
\@nx\ifx\@xa\@nx\csname gmd@detectname@\string#1\endcsname
\@nx\macro@pname
\def\@nx\next{% this \incs{next} will be executed in line
% % \ref{next 3690}.
\SMglobal\RestoreMacro % they both are \incs{protected}.
\@xa\@nx\csname gmd@detect@\string#1\endcsname
\@nx\gmd@charbychar}%^^A we define
% ^^A \inverb|\next| not restore the macro here just in case of
% ^^A multiple declaration of |\def| (in that case there would
% ^^A~be
% ^^A multiple occurences of the macro to be restored in the
% ^^A~carrier
% ^^A \inverb|\gmd@detectors| and we wish
% ^^A all of them not to fire this one time) That's probably too
% ^^A much care since there's the test in line \ref{550}.
\@nx\fi}% of \cs{@nameedef}.
}% of \cs{gmd@adef@selfrestore}.
% The |\ResumeDefining| command takes a~defining command as the
% argument and resumes its automatic detection. Note that it restores
% also the possibly undefined detectors of starred version of the
% argument but that is harmless I~suppose until we have millions of \CSs.
\def\ResumeDefining{\begingroup
\MakePrivateLetters
\gmd@ResumeDfng}
\def\gmd@ResumeDfng#1{%
\escapechar\m@ne
\SMglobal\RestoreMacro*{gmd@detect@\string#1}%
\SMglobal\RestoreMacro*{gmd@detect@\string#1*}%
\endgroup}
% \subdivision{Indexing of \CSs}
% The inner macro indexing macro. |#1| is the |\verb|'s delimiter;
% |#2| is assumed to be the macro's name with MakeIndex-control chars
% quoted. |#3| is a~macro storing the \catother\ macro's name, usually
% |\macro@pname|, built with |\string|ing every char in lines ^^A
% \ref{stringing0}, \ref{stringing1} and \ref{stringing2}. |#3| is
% used only to test if the entry should be specially formatted.
% \changes{v0.98}{06/09/05}{Indexing the codelines improved to sort
% with account of the \cs{filesep} (\cs{HLPrefix})}
\newcommand*\index@macro[3][\verbatimchar]{{%
\@ifundefined{gmd/iexcl/#3}%\label{iexcltest}
{% |#3| is not excluded from index
\@ifundefined{gmd/defentry/#3}% \label{pnametestDef}
{% |#3| is not def entry
\@ifundefined{gmd/usgentry/#3}%\label{pnametestUsg}
{% |#3| is not usg entry
\edef\kind@fentry{\CommonEntryCmd}}%\label{CECmd}
{% |#3| is usg entry
\def\kind@fentry{UsgEntry}%
\un@usgentryze{#3}}%\label{usgentryrs}
}%
{% |#3| is def entry
\def\kind@fentry{DefEntry}%
\un@defentryze{#3}%\label{defentryrs}
}% of |gmd/defentry/| test's `else'
\if@pageindex\@pageinclindexfalse\fi% should it be here or
% there? Definitely here because we'll wish to switch the switch
% with a~declaration.
\if@pageinclindex
\edef\gmu@tempa{gmdindexpagecs{\HLPrefix}{\kind@fentry}{\EntryPrefix}}%
\else
\edef\gmu@tempa{gmdindexrefcs{\HLPrefix}{\kind@fentry}{\EntryPrefix}}%
\fi
\edef\gmu@tempa{\IndexPrefix#2\actualchar%
\quotechar\bslash verb*#1\quoted@eschar#2#1% The last macro in
% this line usually means the first two, but in some cases
% it's redefined to be empty (when we use |\index@macro| to
% index not a~\CS).
\encapchar\gmu@tempa}%
\@xa\special@index\@xa{\gmu@tempa}% We give the
% indexing macro the argument expanded so that \pk{hyperref} may
% see the explicit encapchar in order not to add its own
% encapsulation of \verb+|hyperpage+ when the (default)
% |hyperindex=true|\TextCommonIndex*{hyperindex} option is in
% force. (After this setting the |\edef|s in the above may be
% changed to |\def|s.)
% \changes[\index@macro]{v0.98f}{06/9/30}{explicit MakeIndex
% controls changed to corresponding macros. Therefore
% \cs[]{hyperindex} option of \pk{hyperref} didn't see the
% encapsulation and added its own. So I~expanded the
% argument of the very indexing macro}
}{}% closing of |gmd/iexcl/| test.
}}
\def\un@defentryze#1{%
\@xa\g@relaxen\csname gmd/defentry/#1\endcsname
\ifx\gmd@detectors\empty
\g@relaxen\last@defmark
\fi}% the last macro (assuming \cs{fi} is not a~macro :-)
% is only used by \cs{changes}. If we are in the scope of automatic
% detection of definitions, we want to be able not to use \inverb|\Define|
% but write |\changes| after a~definition and get proper entry. Note
% that in case of automatic detection of definitions
% |\last@defmark|'s value keeps until the next definition.
\def\un@usgentryze#1{%
\@xa\g@relaxen\csname gmd/usgentry/#1\endcsname}
\@emptify\EntryPrefix% this macro seems to be obsolete now
% (v0.98d).
% For the case of page-indexing a~macro in the commentary when
% codeline index option is on:
\newif\if@pageinclindex
\newcommand*\quoted@eschar{\quotechar\bslash}% we'll redefine it when
% indexing an environment.
% Let's initialize |\IndexPrefix|
\def\IndexPrefix{}
% The |\IndexPrefix| and |\HLPrefix| (`HyperLabel Prefix') macros are
% given with account of a~possibility of documenting several files
% in(to) one document. In such case the user may for each file
% |\def\IndexPrefix{|\<package name>|!}| for instance and it will work
% as main level index entry and |\def\HLPrefix{|\<package name>|}|
% as a~prefix in hypertargets in the codelines. They are redefined by
% |\DocInclude| e.g.
\if@linesnotnum\@pageindextrue\fi
\AtBeginDocument{%
\if@pageindex
\def\gmdindexrefcs#1#2#3#4{\csname#2\endcsname{\hyperpage{#4}}}%^^A
% in the page case we gobble the third argument that is supposed
% to be the entry prefix.\ilrr
\let\gmdindexpagecs=\gmdindexrefcs
\else
% \DefIndex\gmdindexrefcs
% \DefIndex\gmdindexpagecs
\def \gmdindexrefcs#1#2#3#4{\gmiflink[clnum.#4]{%
\csname#2\endcsname{#4}}}%
\def \gmdindexpagecs#1#2#3#4{\hyperlink{page.#4}{%
\csname#2\endcsname{\gmd@revprefix{#3}#4}}}%
% \stanza \DefIndex\gmd@revprefix
\def\gmd@revprefix#1{%
\def\gmu@tempa{#1}%
\ifx\gmu@tempa\@empty p.\,\fi}
\providecommand*\HLPrefix{}% it'll be the hypertargets names' prefix
% in mul\-ti-docs. Moreover, it showed that if it was empty,
% \pk{hyperref} saw duplicates of the hyper destinations, which
% was perfectly understandable (|codelinenum.123| made by
% |\refstepcounter| and |codelinenum.123| made by
% |\gmhypertarget|). But since v0.98 it is not a~problem anymore
% because during the automatic \inverb|\hypertarget|ing the
% lines are labeled |clnum.|\<number>. When |\HLPrefix| was
% defined as dot, MakeIndex rejected the entries as `illegal page
% number'. \changes[\HLPrefix]{v0.98a}{06/09/05}{again
% \cs{@empty}fied since \cs{hypertarget}ing the codelines names
% them \cs[]{clnum} (since v0.98).}
\fi}
% The definition is postponed till |\begin{document}| because of the
% |\PageIndex| declaration (added for \docfm-compatibility), see
% line \ref{PageIndex}.
%
% I~design the index to contain hyperlinking numbers whether they are
% the line numbers or page numbers. In both cases the last parameter
% is the number, the one before the last is the name of a~formatting macro
% and in linenumber case the first parameter is a~prefix for proper
% reference in multi-doc.
% I~take account of three kinds of formatting the numbers: 1.~the
% `def' entry, 2.~a~`usage' entry, 3.~a~common entry. As in \docfm , let
% them be underlined, italic and upright respectively.
\def\DefEntry#1{\underline{#1}}
\def\UsgEntry#1{\textit{#1}}
% The third option will be just |\relax| by default:
\def\CommonEntryCmd{relax}
% In line \ref{CECmd} it's |\edef|ed to allow an `unm\oumlaut glich'
% situation that the user wants to have the common index entries
% specially formatted. I~use this to make \emph{all} the index entries
% of the driver part to be `usage', see the source of chapter
% \ref*{Driver}.
% Now let's |\def| the macros declaring a~\CS to be indexed
% special way. Each declaration puts the \catother ed name of the
% macro given it as the argument into proper macro to be |\ifx|ed in
% lines \ref{pnametestDef} and \ref{pnametestUsg} respectively.
% Now we are ready to define a~couple of commands. % The |*| versions of
% them are for marking environments and \emph{implicit} \CSs.
\outer\def\DefIndex{\begingroup
\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\Code@DefIndexStar}{\Code@DefIndex}}
% \changes{v0.98d}{06/9/11}{The macros indexing and marginizing
% macros (and other sequences) made \cs{long}}
%\DefIndex\Code@DefIndex
\long\def\Code@DefIndex#1{\endgroup{%
\escapechar\m@ne% because we will compare the macro's name with
% a~string without the backslash.
\@defentryze{#1}{1}}}
% \label{stringamacro}
% \DefIndex\Code@DefIndexStar
\long\def\Code@DefIndexStar#1{%
\endgroup
\addto@estoindex{#1}%
\@defentryze{#1}{0}}
\def\gmd@justadot{.}
\long\def\@defentryze#1#2{%
\@xa\glet\csname gmd/defentry/\string#1\endcsname\gmd@justadot% The\\
% \LaTeX\ \inverb|\@namedef| macro could not be used since it's not `long'.\ilrr
% \Define\last@defmark
\xdef\last@defmark{\string#1}% we |\string| the argument just in case it's
% a~control sequence. But when it can be a~\CS, we |\@defentryze| in
% a~scope of |\escapechar=-1|, so there will never be a~backslash at
% the beginning of |\last@defmark|'s meaning (unless we
% |\@defentryze| |\\|).
\@xa\gdef\csname gmd/isaCS/\last@defmark\endcsname{#2}}% |#2| is
% either 0 or 1. It is the information whether this entry is a~\CS or
% not.
\long\def\@usgentryze#1{%
\@xa\let\csname gmd/usgentry/\string#1\endcsname\gmd@justadot}
% Initialize |\envirs@toindex|
\@emptify\envirs@toindex
% Now we'll do the same for the `usage' entries:
\outer\def\CodeUsgIndex{\begingroup
\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\Code@UsgIndexStar}{\Code@UsgIndex}}
% The |*| possibility is for marking environments etc.
% \DefIndex\Code@UsgIndex
\long\def\Code@UsgIndex#1{\endgroup{%
\escapechar\m@ne
\global\@usgentryze{#1}}}
% \DefIndex\Code@UsgIndexStar
\long\def\Code@UsgIndexStar#1{%
\endgroup
\addto@estoindex{#1}%
\@usgentryze{#1}}
% For the symmetry, if we want to mark a~control sequence or an
% environment's name to be indexed as a~`normal' entry, let's have:
\outer\def\CodeCommonIndex{\begingroup
\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\Code@CommonIndexStar}{\Code@CommonIndex}}
% \DefIndex\Code@CommonIndex
\long\def\Code@CommonIndex#1{\endgroup}
% \DefIndex\Code@CommonIndexStar
\long\def\Code@CommonIndexStar#1{%
\endgroup\addto@estoindex{#1}}
% And now let's define commands to index the control sequences and
% environments occurring in the narrative.
\long\def\text@indexmacro#1{%
{\escapechar\m@ne \xdef\macro@pname{\xiistring#1}}%
\@xa\quote@mname\macro@pname\relax% we process the \CS's
% name char by char and quote MakeIndex controls. |\relax| is the
% iterating macro's stopper. The scanned \CS's quoted name shall be the
% expansion of |\macro@iname|.
\if\verbatimchar\macro@pname
\def\im@firstpar{[$]}%^^A$
\else\def\im@firstpar{}%
\fi
{\do@properindex% see line \ref{do@properindex}.
\@xa \index@macro\im@firstpar\macro@iname\macro@pname}}
% The macro defined below (and the next one) are executed only before
% a~\catother\ macro's name i.e.\ a~nonempty sequence of \catother\
% character(s). This sequence is delimited (guarded) by |\relax|.
\def\quote@mname{%
\def\macro@iname{}%
\quote@charbychar}
% \DefIndex\quote@charbychar
\def\quote@charbychar#1{%
\if\relax#1% finish quoting when you meet |\relax| or:
\else
\quote@char#1%
\xdef\macro@iname{\macro@iname \gmd@maybequote#1}%
\afterfi\quote@charbychar
\fi}
% The next command will take one argument,
% which in plain version should be a~control sequence and in the
% starred version also a~sequence of chars allowed in environment names or
% made other by |\MakePrivateOthers| macro, taken in the curly braces.
\def\TextUsgIndex{\begingroup
\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\Text@UsgIndexStar}{\Text@UsgIndex}}
% \DefIndex\Text@UsgIndex
\long\def\Text@UsgIndex#1{%
\endgroup\@usgentryze#1%
\text@indexmacro#1}
% \DefIndex\Text@UsgIndexStar
\long\def\Text@UsgIndexStar#1{\endgroup\@usgentryze{#1}%
\text@indexenvir{#1}}
\long\def \text@indexenvir#1{%
\edef\macro@pname{\xiistring#1}%
\if\bslash\@xa\@firstofmany\macro@pname\@@nil% if
% |\string|ed |#1| begins with a~backslash, we will gobble it
% to make MakeIndex not see it.
\edef\gmu@tempa{\@xa\@gobble\macro@pname}%
\@tempswatrue
\else
\let\gmu@tempa\macro@pname
\@tempswafalse
\fi
\@xa\quote@mname\gmu@tempa\relax% \label{quote@mname comm}we
% process |\sting|ed |#1| char by char and quote MakeIndex
% controls. |\relax| is the iterating macro's stopper. The quoted
% |\string|ed |#1| shall be the meaning of |\macro@iname|.
{\if@tempswa
\def\quoted@eschar{\quotechar\bslash}%
\else\@emptify\quoted@eschar\fi% we won't print any backslash before
% an environment's name, but we will before a~\CS's name.
\do@properindex% see line \ref{do@properindex}.
\index@macro\macro@iname\macro@pname}}
\def\TextCommonIndex{\begingroup
\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\Text@CommonIndexStar}{\Text@CommonIndex}}
%\DefIndex\Text@CommonIndex
\long\def\Text@CommonIndex#1{\endgroup
\text@indexmacro#1}
% \DefIndex\Text@CommonIndexStar
\long\def\Text@CommonIndexStar#1{\endgroup
\text@indexenvir{#1}}
% As you see in the lines \ref{defentryrs} and \ref{usgentryrs}, the
% markers of special formatting are reset after first use.
% But we wish the \CSs not only to be indexed special way but also
% to be put in marginpars. So:
\outer\def\CodeMarginize{\begingroup
\MakePrivateLetters
\@ifstarl
{\MakePrivateOthers\egCode@MarginizeEnvir}
{\egCode@MarginizeMacro}}
% One more expansion level because we wish |\Code@MarginizeMacro| not
% to begin with |\endgroup| because in the subsequent macros it's used
% \emph{after} ending the re|\catcode|ing group.
% \DefIndex\egCode@MarginizeMacro
\long\def\egCode@MarginizeMacro#1{\endgroup
\Code@MarginizeMacro#1}
% \DefIndex\Code@MarginizeMacro
\long\def\Code@MarginizeMacro#1{{\escapechar\m@ne
\@xa\glet\csname gmd/2marpar/\string#1\endcsname\gmd@justadot
%^^A \xdef\mname@tomarginpar{\xiistring#1}
}}
% \DefIndex\egCode@MarginizeEnvir
\long\def\egCode@MarginizeEnvir#1{\endgroup
\Code@MarginizeEnvir{#1}}
% \DefIndex\Code@MarginizeEnvir
\long\def\Code@MarginizeEnvir#1{\addto@estomarginpar{#1}}
% And a~macro really putting the environment's name in a~marginpar
% shall be trigged at the beginning of the nearest codeline.
%
% Here it is:
\def\mark@envir{%\label{mark@envir}
\ifx\envirs@tomarginpar\@empty
\else
\let\do\Text@Marginize
\envirs@tomarginpar%
\g@emptify\envirs@tomarginpar%
\fi
\ifx\envirs@toindex\@empty
\else
\gmd@doindexingtext
\envirs@toindex
\g@emptify\envirs@toindex%
\fi}
\def\gmd@doindexingtext{%
\def\do##1{% the |\envirs@toindex| list contains |\string|ed
% macros or environments' names in braces and each preceded
% with |\do|. We extract the definition because we use it also in
% line \ref{gmd@doindexingtext 2nd use}.
\if\bslash\@firstofmany##1\@@nil% if
% |##1| begins with a~backslash, we will gobble it for
% MakeIndex not see it.
\edef\gmd@resa{\@gobble##1}%
\@tempswatrue
\else
\edef\gmd@resa{##1}\@tempswafalse
\fi
\@xa\quote@mname\gmd@resa\relax% see line
% \ref{quote@mname comm} \& subs. for commentary.
{\if@tempswa
\def\quoted@eschar{\quotechar\bslash}%
\else\@emptify\quoted@eschar\fi
\index@macro\macro@iname{##1}}}%
}
% One very important thing: initialisation of the list macros:
\@emptify\envirs@tomarginpar
\@emptify\envirs@toindex
% For convenience we'll make the `private letters' first not to bother
% ourselves with |\makeatletter| for instance when we want mark some
% \CS. And |\MakePrivateOthers| for the environment and other string
% case.
\outer\def\Define{\begingroup
\MakePrivateLetters
% We do |\MakePrivateLetters| before |\@ifstarl| in order to avoid
% a~situation that \TeX\ sees a~control sequence with improper name
% (another \CS than we wished)
% (because |\@ifstarl| establishes the |\catcode|s for the next token):
\@ifstarl{\MakePrivateOthers\Code@DefEnvir}{\Code@DefMacro}}
\outer\def\CodeUsage{\begingroup
\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\Code@UsgEnvir}{\Code@UsgMacro}}
% And then we launch the macros that close the group and do the work.
\long\def\Code@DefMacro#1{%
\Code@DefIndex#1% we use the internal macro; it'll close the group.
\Code@MarginizeMacro#1}
% \DefIndex\Code@UsgMacro
\long\def\Code@UsgMacro#1{%
\Code@UsgIndex#1% here also the internal macro; it'll close the group
\Code@MarginizeMacro#1}
% The next macro is taken verbatim ;-) from \docfm\ and the subsequent
% |\let|s, too.
% \DefIndex\codeline@wrindex
\def\codeline@wrindex#1{\if@filesw
\immediate\write\@indexfile
{\string\indexentry{#1}%
{\HLPrefix\number\c@codelinenum}}\fi}
\def\codeline@glossary#1{% It doesn't need to establish a~group since
% it is always called in a~group.
\if@pageinclindex
\edef\gmu@tempa{gmdindexpagecs{\HLPrefix}{relax}{\EntryPrefix}}%
\else
\edef\gmu@tempa{gmdindexrefcs{\HLPrefix}{relax}{\EntryPrefix}}% \inverb|relax| stands for the formatting command. But we don't want to do anything special with the change history entries. \ilrr
\fi
\protected@edef\gmu@tempa{%
\@nx\protected@write\@nx\@glossaryfile{}%
{\string\glossaryentry{#1\encapchar\gmu@tempa}%
{\HLPrefix\number\c@codelinenum}}}%
\gmu@tempa
}
% \changes[\changes]{v0.99m}{2008/08/09}{changed to write the line
% number instead of page number by default and with
% \env{codelineindex} option which seems to be more reasonable
% especially with the \env{countalllines} option}
% We initialize it due to the option (or lack of the option):
\AtBeginDocument{%
\if@pageindex
\let\special@index=\index
\let\gmd@glossary\glossary
\else
% \DefIndex\special@index
\let\special@index=\codeline@wrindex
\let\gmd@glossary\codeline@glossary
% \label{codeline indexnumber declaration}
\fi}% postponed till |\begin{document}| with respect of \docfm-like
% declarations.
% And in case we don't want to index:
\def\gag@index{\let\index=\@gobble
%\label{gag@index}
\let\codeline@wrindex=\@gobble}
% We'll use it in one more place or two. And we'll wish to be able to
% undo it so
% let's copy the original meanings:
% \DefIndex\@@index \DefIndex\@@codeline@wrindex
\StoreMacros{\index\codeline@wrindex}
\def\ungag@index{\RestoreMacros{\index\@@codeline@wrindex}}
%\label{ungag@index}
% Our next task is to define macros that'll mark and index an
% environment or other string in the code. Because of lack of
% a~backslash, no environment's name is scanned so we have to proceed
% different way. But we wish the user to have symmetric tools, i.e.,
% the `def' or `usage' use of an environment should be declared before
% the line where the environment occurs. Note the slight difference
% between these and the commands to declare a~\CS marking: the latter
% do not require to be used \emph{immediately} before the line containig the
% \CS to be marked. We separate indexing from marginizing to leave
% a~possibility of doing only one of those things.
% \DefIndex\Code@DefEnvir
\long\def\Code@DefEnvir#1{%
\endgroup
\addto@estomarginpar{#1}%
\addto@estoindex{#1}%
\@defentryze{#1}{0}}
% \DefIndex\Code@UsgEnvir
\long\def\Code@UsgEnvir#1{%
\endgroup
\addto@estomarginpar{#1}%
\addto@estoindex{#1}%
\@usgentryze{#1}}
% \DefIndex\addto@estomarginpar
\long\def\addto@estomarginpar#1{%
\edef\gmu@tempa{\@nx\do{\xiistring#1}}% we |\string| the argument to
% allow it to be a~control sequence.
\@xa\addtomacro\@xa\envirs@tomarginpar\@xa{\gmu@tempa}}
% \DefIndex\addto@estoindex
\long\def\addto@estoindex#1{%
\edef\gmu@tempa{\@nx\do{\xiistring#1}}
\@xa\addtomacro\@xa\envirs@toindex\@xa{\gmu@tempa}}
% And now a~command to mark a~`usage' occurrence of a~\CS, environment
% or another string in the commentary. As the `code' commands this also
% has plain and starred version, first for \CSs appearing explicitly and the
% latter for the strings and \CSs appearing implicitly.
\def\TextUsage{\begingroup
%\label{TextUsage}
\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\Text@UsgEnvir}{\Text@UsgMacro}}
% \DefIndex\Text@UsgMacro
\long\def\Text@UsgMacro#1{%
\endgroup{\tt\xiistring#1}%
\Text@Marginize#1%
\begingroup\Code@UsgIndex#1% we declare the kind of formatting of the entry.
\text@indexmacro#1}
% \DefIndex\Text@UsgEnvir
\long\def\Text@UsgEnvir#1{%
\endgroup{\tt\xiistring#1}%
\Text@Marginize{#1}%
\@usgentryze{#1}% we declare the `usage' kind of formatting of the
% entry and index the sequence |#1|.
\text@indexenvir{#1}}
% We don't provide commands to mark a~macro's or environment's
% definition present within the narrative because we think there won't
% be any: one defines macros and environments in the code not in
% the commentary.
\def\TextMarginize{\begingroup
\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\egText@Marginize}{\egText@Marginize}}
% \DefIndex\egText@Marginize
\long\def\egText@Marginize#1{\endgroup
\Text@Marginize#1}
% We check whether the margin pars are enabled and proceed
% respectively in either case.
\if@marginparsused
\reversemarginpar
\marginparpush\z@
\marginparwidth8pc\relax
% ^^A \settowidth{\marginparsep}{\ \ }%
%
% You may wish to put not only macros and environments to
% a~marginpar.
\long\def\gmdmarginpar#1{%
\marginpar{\raggedleft\strut
\hskip0ptplus100ptminus100pt%
#1}}%\stanza
%
\else
\long\def\gmdmarginpar#1{}%
\fi
\long\def\Text@Marginize#1{%
\gmdmarginpar{\marginpartt\xiistring#1}}
% Note that the above macro will just gobble its argument if the
% marginpars are disabled.^^A\NoEOF so far O.K.
%
% It may be advisable to choose a~condensed typewriter font for the
% marginpars, if there is any. (The Latin Modern font family provides
% a~light condensed typewriter font, it's set in \pk{gmdocc} class.)
\let\marginpartt\tt
% If we pront also the narration lines' numbers, then the index entries for \CSs and
% environments marked in the commentary should have codeline numbers
% not page numbers and that is |\let| in line
% \ref{codeline indexnumber declaration}. On the other hand, if we
% don't print narration lines' numbers, then a~macro or an environment marked in
% the commentary should have page number not codeline number. This we
% declare here, among others\ we add the letter |p| before the page number.
% \DefIndex\do@properindex
\def\do@properindex{%\label{do@properindex}
\if@printalllinenos\else
\@pageinclindextrue
\let\special@index=\index
\fi}
% \stanza
% In \docfm\ all the `working' \TeX\ code should be braced in(to) the
% \env{macrocode} environments. Here another solutions are taken so to
% be \docfm-compatible we only should
% nearly-ignore \env{macrocode(*)}s
% with their Percent and The Four Spaces Preceding ;-)\,. I.e., to ensure
% the line ends are `queer'. And that the \ds\ directives will be
% typeset as the \ds\ directives. And that the usual code escape
% char will be restored at |\end{macrocode}|. And to add the vertical
% spaces.
%
% ^^A~\NoEOF so far \acro{OK}
% If you know \docfm\ conventions, note that \gmdoc\ \emph{does not}
% require |\end{macrocode}| to be preceded ^^A(
% with any particular number of any char :-)\,.
% \changes{v0.98b}{06/09/07}{To \env{macrocode(*)} default
% definitions \cs{QueerEOL} added}
\newenvironment*{macrocode*}{%\label{macrocode*}
\if@codeskipput\else\par\addvspace\CodeTopsep\@codeskipputgtrue\fi
\QueerEOL}%^^A\docstrips
{\par\addvspace\CodeTopsep\CodeEscapeChar\\}
% Let's remind that the starred version makes | |
% visible, which is the default in \pk{gmdoc} outside
% \env{macrocode}.
%
% So we should make the spaces \emph{invisible} for the unstarred
% version.
\newenvironment*{macrocode}{%\label{macrocode}
\if@codeskipput\else\par\addvspace\CodeTopsep\@codeskipputgtrue\fi
\QueerEOL}% \changes{v0.99l}{2008/08/06}{removed \cs{CodeSpacesBlank}}
{\par\addvspace\CodeTopsep\CodeEscapeChar\\}
% Note that at the end of both the above environments the |\|'s
% r\ocircum le
% as the code escape char is restored. This is crafted for the
% |\SpecialEscapechar| macro's compatibility: this macro influences
% only the first \env{macrocode} environment. The situation that the
% user wants some queer escape char in general and in a~particular
% \env{macrocode} yet another seems to me ``unm\oumlaut glich,
% Prinzessin''\footnote{Richard Strauss after Oscar Wilde, ^^B
% \textit{Salome}.}.
%
% \stanza
% Since the first \file{.dtx} I~tried to compile after the first
% published version of \gmdoc\ uses a~lot of commented out code in
% \env{macrocode}s, it seems to me necessary to add a~possibility to
% typeset \env{macrocode}s as if they were a~kind of \env{verbatim},
% that is to leave the code layer and narration layer philosophy.
%
%\Define*{oldmc}
\let\oldmc\macrocode
\let\endoldmc\endmacrocode
%\Define*{oldmc*}
\n@melet{oldmc*}{macrocode*}
\n@melet{endoldmc*}{endmacrocode*}
% Now we arm \env{oldmc} and \env{olmc*} with
% the macro looking for |% \end|\arg{envir name}.
\addtomacro\oldmc{\@oldmacrocode@launch}%
\@xa\addtomacro\csname oldmc*\endcsname{%
\@oldmacrocode@launch}
% \DefIndex\@oldmacrocode@launch
\def\@oldmacrocode@launch{%
\emptify\gmd@textEOL% to disable it in |\gmd@docstripdirective|
% launched within the code.
\gmd@ctallsetup
\glet\stored@code@delim\code@delim
\@makeother\^^B\CodeDelim\^^B%
\ttverbatim \gmd@DoTeXCodeSpace%
\@makeother\|% because |\ttverbatim| doesn't do that.
\MakePrivateLetters% see line \ref{MPL}.\par
% ^^A \@xa\@makeother\code@delim
\docstrips@percent \@makeother\>%
% sine qua non of the automatic delimiting is replacing possible
% |*|\catother in the environment's name with |*|\catletter.
% Not to complicate assume |*| may occur at most once and only at
% the end. We also assume the environment's name consists only of
% character tokens whose catcodes (except of |*|) will be the same
% in the verbatim text.
\@xa\gmd@currenvxistar\@currenvir*\relax
\@oldmacrocode}
\foone{\catcode`*11 }
{\def\gm@xistar{*}}
\def\gmd@currenvxistar#1*#2\relax{%
\edef\@currenvir{#1\if*#2\gm@xistar\fi}}
% The trick is that |#2| may be either |*|\catother\ or empty. If it's
% |*|, the test is satisfied and |\if...\fi| expands to
% |\gm@xistar|. If |#2| is empty, the test is also satisfied since
% |\gm@xistar| expands to |*| but there's nothing to expand to. So, if
% the environment's name ends with |*|\catother, it'll be substituted
% with |*|\catletter or else nothing will be added. (Note that a~|*|
% not at the end of env.\ name would cause a~disaster.)
%^^A~ \Define\@oldmacrocode
\foone{%
\catcode`[=1 \catcode`]=2
\catcode`\{=\active \@makeother\}
\@makeother\^^B
\catcode`/=0 \catcode`\\=\active
\catcode`&=14 \catcode`*=11
\catcode`\%=\active \obeyspaces}& % \CodeEscapeChar\/ \CDAnd
[& here the \cs{foone}'s second pseudo-argument begins
&%\stanza
/def/@oldmacrocode[&
/bgroup/let =/relax&% to avoid writing |/@nx | four times.
/xdef/oldmc@def[&
/def/@nx/oldmc@end####1/@nx% /@nx\end&
/@nx{/@currenvir}[&
####1^^B/@nx/end[/@currenvir]/@nx/gmd@oldmcfinis]]&
/egroup&% now |\oldmc@edef| is defined to have one parameter delimited
&% with |\end|\arg{current env.'s name}
/oldmc@def&
/oldmc@end]&% \CDPerc
]
\def\gmd@oldmcfinis{%
\@xa\CodeDelim\stored@code@delim
\gmd@mchook}% see line \ref{gmd@mchook}
\def\OldMacrocodes{%% \changes{v0.99m}{2008/08/10}{renamed from
%% \cs{VerbMacrocodes}}
\let\macrocode\oldmc
\n@melet{macrocode*}{oldmc*}}
% To handle \ds\ directives in the code (in the old macrocodes
% case that is).
% \begin{oldmc}
\foone{\catcode`\%\active}
{\def\docstrips@percent{\catcode`\%\active
\let%\gmd@codecheckifds}}
% \end{oldmc}\
% The point is, the active |%| will be expanded when just after it is
% the |\gmd@charbychar| cs token and next is some char, the |^^B| code
% delimiter at least. So, if that char is |<|, we wish to launch
% \ds\ directive typesetting. (Thanks to |\ttverbatim| all the |<|
% are `other'.)
\def\gmd@codecheckifds#1#2{% note that |#1| is just to gobble
% \inverb|\gmd@charbychar| token.
% ^^A \typeout{code if ds \on@line}%
\if@dsdir\@dsdirgfalse
\if\@nx<\@nx#2\afterfifi\gmd@docstripdirective
\else\afterfifi{\xiipercent#1#2}%
\fi
\else\afterfi{\xiipercent#1#2}%
\fi}
%
% \begin{environment}{macro}
% Almost the same we do with the \env{macro(*)} environments, stating
% only their argument to be processed as the `def' entry. Of course,
% we should re|\catcode| it first.
% \DefIndex*\macro
\newenvironment{macro}{%\label{macro}
\@tempskipa=\MacroTopsep
\if@codeskipput\advance\@tempskipa by-\CodeTopsep\fi
\par\addvspace{\@tempskipa}\@codeskipputgtrue
\begingroup\MakePrivateLetters\MakePrivateOthers% we make also the
% `private others' to cover the case of other sequence in the
% argument. (We'll use the |\macro| macro also in the environment
% for describing and defining environments.)
\gmd@ifonetoken\Hybrid@DefMacro\Hybrid@DefEnvir}%
% \DefIndex*\endmacro
{\par\addvspace\MacroTopsep\@codeskipputgtrue}
% It came out that the \docfm's author(s) give the \env{macro}
% environment also starred versions of commands as argument. It's \acro{OK}
% since (the default version of) |\MakePrivateLetters| makes |*|
% a~letter and therefore such a~starred version is just one \CS.
% However, in \pk{doc.dtx} occur \env{macro}s that mark
% \emph{implicit} definitions i.e., such that the defined \CS is not
% scanned in the subsequent code.
%
% \begin{macro*}{macro*}
% And for those who want to to use this environment
% for marking implicit definitions, define the star
% version:
\@namedef{macro*}{\let\gmd@ifonetoken\@secondoftwo\macro}
% \DefIndex*\endmacro*
\@xa\let\csname endmacro*\endcsname\endmacro
% Note that \env{macro} and \env{macro*} have the same effect for
% more-than-one-token arguments thanks to |\gmd@ifonetoken|'s meaning
% inside unstarred \env{macro} (it checks whether the argument is
% one-token and if it isn't, |\gmd@ifonetoken| switches execution to
% `other sequence' path).
%
% The two environments behave different only with a~one-token
% argument: \env{macro} postpones indexing it till the first scanned
% occurrence while \env{macro*} till the first code line met.
% \end{macro*}
% \end{environment}
%
% \stanza
% Now, let's complete the details.
% First define an |\if|-like macro that turns true when the string
% given to it consists of just one token (or one |{|\<text>|}|, to
% tell the whole truth).
\def\gmd@ifsingle#1#2\@@nil{%
\def\gmu@tempa{#2}%
\ifx\gmu@tempa\@empty}
% Note it expands to an open |\if...| test (unbalanced with |\fi|) so it
% has to be used as all the |\if|s, with optional |\else| and
% obligatory |\fi|. And cannot be used in the possibly skipped
% branches of other |\if...|s (then it would result with `extra
% |\fi|/extra |\else|' errors). But the below usage is safe since both
% |\gmd@ifsingle| and its |\else| and |\fi| are hidden in a~macro
% (that will not be |\expandafter|ed).
%
% Note also that giving |\gmd@ifsingle| an |\if...| or so as the first
% token of the argument will not confuse \TeX\ since the first token
% is just gobbled. The possibility of occurrence of |\if...| or so as
% a~not-first token seems to be negligible.
\def\gmd@ifonetoken#1#2#3{%
\def\gmu@tempb{#3}% We hide |#3| from \TeX\ in case it's |\if...| or
% so. \inverb|\gmu@tempa| is used in \inverb|\gmd@ifsingle|. \ilrr
\gmd@ifsingle#3\@@nil
\afterfi{\@xa#1\gmu@tempb}%
\else
\edef\gmu@tempa{\@xa\string\gmu@tempb}%
\afterfi{\@xa#2\@xa{\gmu@tempa}}%
\fi}
% Now, define the mysterious |\Hybrid@DefMacro| and |\Hybrid@DefEnvir|
% macros. They mark their argument with a~certain subtlety: they put
% it in a~marginpar at the point where they are and postpone indexing
% it till the first scanned occurrence or just the first code line met.
\long\def\Hybrid@DefMacro#1{%
\Code@DefIndex{#1}% this macro closes the group opened by |\macro|.
\Text@MarginizeNext{#1}}
\long\def\Hybrid@DefEnvir#1{%
\Code@DefIndexStar{#1}% this macro also closes the group begun by
% |\macro|.
\Text@MarginizeNext{#1}}
\long\def\Text@MarginizeNext#1{%
\gmd@evpaddonce{\Text@Marginize{#1}\ignorespaces}}
%^^A~\strut from before Text@Marginize deleted as unnecessary in everypar.
% The following macro adds its argument to |\everypar| using an auxiliary
% macro to wrap the stuff in. The auxiliary macro has
% a~self-destructor built in so it |\relax|es itself after first use.
\long\def\gmd@evpaddonce#1{%
\global\advance\gmd@oncenum\@ne
\@xa\long\@xa\edef% \CodeUsgIndex*{NeuroOncer}
\csname gmd/evp/NeuroOncer\the\gmd@oncenum\endcsname{%
\@nx\g@relaxen
\csname gmd/evp/NeuroOncer\the\gmd@oncenum\endcsname}% Why does it
% work despite it shouldn't? Because when the \CS got with
% \inverb|\csname...\endcsname| is undefined, it's equivalent
% |\relax| and therefore unexpandable. That's why it passes
% \inverb|\edef| and is able to be assigned.\ilrr
\@xa\addtomacro\csname gmd/evp/NeuroOncer\the\gmd@oncenum\endcsname{#1}%
\@xa\addto@hook\@xa\everypar\@xa{%
\csname gmd/evp/NeuroOncer\the\gmd@oncenum\endcsname}%
}
\newcount\gmd@oncenum
%^^A % We store the number uniquifying the auxiliary
%^^A % macro in a~macro to save count registers (cf.\ \pk{gmutils}
%^^A % sec.\ \gmiflink{To Save Precious Count Registers}).
% \begin{environment}{environment}
% Wrapping a~description and definition of an environment in
% a~\env{macro} environment would look inappropriate (`zgrzyta\l o by'
% in Polish) although there's no \TeX nical obstacle to do
% so. Therefore we define the \env{environment}, because of \ae^^B
% sthetic and psychological reasons.
\@xa\let\@xa\environment\csname macro*\endcsname
\@xa\let\@xa\endenvironment\csname endmacro*\endcsname
% \end{environment}
%
%\subdivision[Index exclude list]{\gmhypertarget{Index exclude list}}
% We want some \CSs not to be indexed, e.g., the \LaTeX\
% internals and \TeX\ primitives.
%
% \docfm\ takes |\index@excludelist| to be a~|\toks| register to store
% the list of expelled \CSs. Here we'll deal another way. For each \CS
% to be excluded we'll make (|\let|, to be precise) a~control sequence
% and then we'll be checking if it's undefined (|\ifx|-equivalent
% |\relax|).\footnote{This idea comes from Marcin Woli\nacute ski.}
%
\def\DoNotIndex{\bgroup\MakePrivateLetters\DoNot@Index}
% \changes{v0.97}{06/09/04}{\cs{MakePrivateLetters} added}
% \changes{v0.98}{06/09/05}{Removed since it had not worked. The
% macros in the argument should be separated with commas. I~understood
% why it didn't work: because 't was iside a~command. So I~put it back}
% \changes{v0.98d}{06/9/11}{Unmade \cs{global}}
\long\def\DoNot@Index#1{\egroup% we close the group,
\let\gmd@iedir\gmd@justadot% we declare the direction of the cluding
% to be \emph{ex}cluding. We act this way to be able to reverse the
% exclusions easily later.\ilrr
\dont@index#1.}
% \DefIndex\dont@index
\long\def\dont@index#1{%
\def\gmu@tempa{\@nx#1}% My \TeX\ Guru's trick to deal with |\fi|
% and such, i.e., to hide from \TeX\ when it is processing a~test's
% branch without expanding.
\if\gmu@tempa.% a~dot finishes expelling
\else
\if\gmu@tempa,% The list this macro is put before may
% contain commas and that's O.K., we just continue the work.
\afterfifi\dont@index
\else% what is else shall off the Index be expelled.
{\escapechar\m@ne
\xdef\gmu@tempa{\string#1}}%
\@xa\let%
\csname gmd/iexcl/\gmu@tempa\endcsname=\gmd@iedir% In the default
% case explained e.g.\ by the macro's name, the last macro's
% meaning is such that the test in line \ref{iexcltest} will
% turn false and the subject \CS shall not be indexed.
% We |\let| not |\def| to spare \TeX's memory.
\afterfifi\dont@index
\fi
\fi}
%^^A(
% Let's now give the exclude list copied \*verbatim ;-) from
% \pk{doc.dtx}. I~give it in the code layer because I~suppose one will
% document not \LaTeX\ source but normal packages.
%
% \begin{CodeSpacesSmall}
% \DoNotIndex\DoNotIndex \Define\DefaultIndexExclusions
\DoNotIndex\{ \DoNotIndex\}% \label{DNIbraces}the index entries of
% these two \CSs would be rejected by MakeIndex anyway.
\begin{MakePrivateLetters}% Yes, |\DoNotIndex| does
% |\MakePrivateLetters| on its own but No, it won't have any effect
% if it's given in another macro's |\def|.
% \HideAllDefining
\gdef\DefaultIndexExclusions{%
\DoNotIndex{\@ \@@par \@beginparpenalty \@empty}%\label{DIE1}
\DoNotIndex{\@flushglue \@gobble \@input}%
\DoNotIndex{\@makefnmark \@makeother \@maketitle}%
\DoNotIndex{\@namedef \@ne \@spaces \@tempa}%
\DoNotIndex{\@tempb \@tempswafalse \@tempswatrue}%
\DoNotIndex{\@thanks \@thefnmark \@topnum}%
\DoNotIndex{\@@ \@elt \@forloop \@fortmp \@gtempa \@totalleftmargin}%
\DoNotIndex{\" \/ \@ifundefined \@nil \@verbatim \@vobeyspaces}%
\DoNotIndex{\| \~ \ \active \advance \aftergroup \begingroup \bgroup}%
\DoNotIndex{\mathcal \csname \def \documentstyle \dospecials \edef}%
\DoNotIndex{\egroup}%
\DoNotIndex{\else \endcsname \endgroup \endinput \endtrivlist}%
\DoNotIndex{\expandafter \fi \fnsymbol \futurelet \gdef \global}%
\DoNotIndex{\hbox \hss \if \if@inlabel \if@tempswa \if@twocolumn}%
\DoNotIndex{\ifcase}%
\DoNotIndex{\ifcat \iffalse \ifx \ignorespaces \index \input \item}%
\DoNotIndex{\jobname \kern \leavevmode \leftskip \let \llap \lower}%
\DoNotIndex{\m@ne \next \newpage \nobreak \noexpand \nonfrenchspacing}%
\DoNotIndex{\obeylines \or \protect \raggedleft \rightskip \rm \sc}%
\DoNotIndex{\setbox \setcounter \small \space \string \strut}%
\DoNotIndex{\strutbox}%
\DoNotIndex{\thefootnote \thispagestyle \topmargin \trivlist \tt}%
\DoNotIndex{\twocolumn \typeout \vss \vtop \xdef \z@}%
\DoNotIndex{\, \@bsphack \@esphack \@noligs \@vobeyspaces \@xverbatim}%
\DoNotIndex{\` \catcode \end \escapechar \frenchspacing \glossary}%
\DoNotIndex{\hangindent \hfil \hfill \hskip \hspace \ht \it \langle}%
\DoNotIndex{\leaders \long \makelabel \marginpar \markboth \mathcode}%
\DoNotIndex{\mathsurround \mbox}%\verb+% \newcount \newdimen \newskip+
\DoNotIndex{\nopagebreak}%
\DoNotIndex{\parfillskip \parindent \parskip \penalty \raise \rangle}%
\DoNotIndex{\section \setlength \TeX \topsep \underline \unskip}%
\DoNotIndex{\vskip \vspace \widetilde \\ \% \@date \@defpar}%
\DoNotIndex{\[ \]}% see line \ref{DNIbraces}.
\DoNotIndex{\count@ \ifnum \loop \today \uppercase \uccode}%
\DoNotIndex{\baselineskip \begin \tw@}%
\DoNotIndex{\a \b \c \d \e \f \g \h \i \j \k \l \m \n \o \p \q}%
\DoNotIndex{\r \s \t \u \v \w \x \y \z \A \B \C \D \E \F \G \H}%
\DoNotIndex{\I \J \K \L \M \N \O \P \Q \R \S \T \U \V \W \X \Y \Z}%
\DoNotIndex{\1 \2 \3 \4 \5 \6 \7 \8 \9 \0}%
\DoNotIndex{\! \# \$ \& \' \( \) \. \: \; \< \= \> \? \_}% |\+|
% seems to be so rarely used that it may be advisable to index it.
\DoNotIndex{\discretionary \immediate \makeatletter \makeatother}%
\DoNotIndex{\meaning \newenvironment \par \relax \renewenvironment}%
\DoNotIndex{\repeat \scriptsize \selectfont \the \undefined}%
\DoNotIndex{\arabic \do \makeindex \null \number \show \write \@ehc}%
\DoNotIndex{\@author \@ehc \@ifstar \@sanitize \@title}%
\DoNotIndex{\if@minipage \if@restonecol \ifeof \ifmmode}%
\DoNotIndex{\lccode %|%\newtoks|
\onecolumn \openin \p@ \SelfDocumenting}%
\DoNotIndex{\settowidth \@resetonecoltrue \@resetonecolfalse \bf}%
\DoNotIndex{\clearpage \closein \lowercase \@inlabelfalse}%
\DoNotIndex{\selectfont \mathcode \newmathalphabet \rmdefault}%
\DoNotIndex{\bfdefault}%
% From the above list I~removed some |\new...| declarations because
% I~think it may be useful to see gathered the special |\new...|s of
% each kind. For the same reason I~would not recommend excluding
% from the index such declarations as |\AtBeginDocument|,
% |\AtEndDocument|, |\AtEndOfPackage|, |\DeclareOption|,
% |\DeclareRobustCommand| etc. But the
% common definitions, such as |\new/providecommand| and
% |\(e/g/x)def|s, as the most common, in my opinion excluded should
% be.^^A\)
%
% \stanza
% And some my exclusions:
\DoNotIndex{\@@input \@auxout \@currentlabel \@dblarg}%
\DoNotIndex{\@ifdefinable \@ifnextchar \@ifpackageloaded}%
\DoNotIndex{\@indexfile \@let@token \@sptoken \^}% the latter comes
% from \CSs like |\^^M|, see sec. \ref{Tasks}.
\DoNotIndex{\addto@hook \addvspace}%
\DoNotIndex{\CurrentOption}%
\DoNotIndex{\emph \empty \firstofone}%
\DoNotIndex{\font \fontdimen \hangindent \hangafter}%
\DoNotIndex{\hyperpage \hyperlink \hypertarget}%
\DoNotIndex{\ifdim \ifhmode \iftrue \ifvmode \medskipamount}%
\DoNotIndex{\message}%
\DoNotIndex{\NeedsTeXFormat \newcommand \newif}%
\DoNotIndex{\newlabel}%
\DoNotIndex{\of}%
% ^^A\PassOptionsToClass\PassOptionsToPackage
\DoNotIndex{\phantom \ProcessOptions \protected@edef}%
\DoNotIndex{\protected@xdef \protected@write}%
\DoNotIndex{\ProvidesPackage \providecommand}%
\DoNotIndex{\raggedright}%
\DoNotIndex{\raisebox \refstepcounter \ref \rlap}% ^^A\RequirePackage
\DoNotIndex{\reserved@a \reserved@b \reserved@c \reserved@d}%
\DoNotIndex{\stepcounter \subsection \textit \textsf \thepage \tiny}%
\DoNotIndex{\copyright \footnote \label \LaTeX}%
% \changes[\DefaultIndexExclusions]{v0.98a}{06/09/06}{more
% macros added to the `exclude list'}
\DoNotIndex{\@eha \@endparenv \if@endpe \@endpefalse \@endpetrue}%
\DoNotIndex{\@evenfoot \@oddfoot \@firstoftwo \@secondoftwo}%
\DoNotIndex{\@for \@gobbletwo \@idxitem \@ifclassloaded}%
\DoNotIndex{\@ignorefalse \@ignoretrue \if@ignore}%
\DoNotIndex{\@input@ \@input}%
\DoNotIndex{\@latex@error \@mainaux \@nameuse}%
\DoNotIndex{\@nomath \@oddfoot}%|%\@onlypreamble| should be indexed
% \acro{IMO}.
\DoNotIndex{\@outerparskip \@partaux \@partlist \@plus}%
\DoNotIndex{\@sverb \@sxverbatim}%
\DoNotIndex{\@tempcnta \@tempcntb \@tempskipa \@tempskipb}%\\
% I~think the layout parameters even the kernel, should not be
% excluded: \inverb|\@topsep| \inverb|\@topsepadd|^^B
% \ \inverb|\abovedisplayskip| \inverb|\clubpenalty| etc.
\DoNotIndex{\@writeckpt}%
\DoNotIndex{\bfseries \chapter \part \section \subsection}%
\DoNotIndex{\subsubsection}%
\DoNotIndex{\char \check@mathfonts \closeout}%
\DoNotIndex{\fontsize \footnotemark \footnotetext \footnotesize}%
\DoNotIndex{\g@addto@macro \hfilneg \Huge \huge}%
\DoNotIndex{\hyphenchar \if@partsw \IfFileExists }%
\DoNotIndex{\include \includeonly \indexspace}%
\DoNotIndex{\itshape \language \LARGE \Large \large}%
\DoNotIndex{\lastbox \lastskip \m@th \makeglossary}%
\DoNotIndex{\maketitle \math@fontsfalse \math@fontstrue \mathsf}%
\DoNotIndex{\MessageBreak \noindent \normalfont \normalsize}%
\DoNotIndex{\on@line \openout \outer}%
\DoNotIndex{\parbox \part \rmfamily \rule \sbox}%
\DoNotIndex{\sf@size \sffamily \skip}%
\DoNotIndex{\textsc \textup \toks@ \ttfamily \vbox}%
% \changes[\DoNotIndex]{v0.97}{06/09/04}{Excluding some star-versions of
% commands}
% \nostanza\noindent
% |%% \DoNotIndex{\begin*}| maybe in the future, if the idea gets
% popular\dots \nostanza
\DoNotIndex{\hspace* \newcommand* \newenvironment* \providecommand*}%
\DoNotIndex{\renewenvironment* \section* \chapter*}%\label{DIE2}
}% of |\DefaultIndexExclusions|.\par
% I~put all the expellings into a~macro because I~want them
% to be optional.
\end{MakePrivateLetters}
% \end{CodeSpacesSmall} \ResumeAllDefining
% And we execute it due to the (lack of) counter-corresponding option:
\if@indexallmacros\else
\DefaultIndexExclusions
\fi
% If we expelled so many \CSs, someone may like it in general but
% he/she may need one or two expelled to be indexed back. So
%
\def\DoIndex{\bgroup\MakePrivateLetters\Do@Index}
% \changes{v0.97}{06/09/04}{\cs{MakePrivateLetters} added.}
% \changes{v0.98}{06/09/05}{\cs{MakePrivateLetters} removed since it
% hadn't worked. I~understood why it didn't work: because
% 't was iside a~command and I~put it back}
% \changes{v0.98d}{06/9/11}{Unmade \cs{global}}
\long\def\Do@Index#1{\egroup\@relaxen\gmd@iedir\dont@index#1.}% note
% we only redefine an auxiliary \CS and launch also |\dont@index| inner
% macro.
% And if a~user wants here make default exclusions and there do not
% make them, \heshe\ may use the |\DefaultIndexExclusions| declaration
% \himher self. This declaration \acro{OCSR}, but anyway let's provide the
% counterpart. It \acro{OCSR}, too.
%
\def\UndoDefaultIndexExclusions{%
\StoreMacro\DoNotIndex
% ^^A\dont@index
\let\DoNotIndex\DoIndex
% ^^A \let\dont@index=\do@index
\DefaultIndexExclusions
% ^^A \RestoreMacro\dont@index
\RestoreMacro\DoNotIndex}
% \changes{v0.98d}{06/9/11}{Unmade \cs{global}}
%
%
% \subdivision{Index parameters}
%
% \begin{quotation}The |\IndexPrologue| macro is used to place a short
% message into the document above the index. It is implemented by
% redefining |\index@prologue|, a macro which holds the default
% text. We'd better make it a |\long| macro to allow |\par| commands
% in its argument.\end{quotation}
\long\def\IndexPrologue#1{\@bsphack\def\index@prologue{#1}\@esphack}
% \label{IndexPrologue}
\def\indexdiv{\@ifundefined{chapter}{\section*}{\chapter*}}
% \changes{v0.98a}{06/09/06}{modified to contain the star so that
% it may be redefined for tests.}
\@ifundefined{index@prologue} {\def\index@prologue{\indexdiv{Index}%
\markboth{Index}{Index}%
Numbers written in italic refer to the \if@pageindex pages \else
code lines \fi where the
corresponding entry is described; numbers underlined refer to the
\if@pageindex\else code line of the \fi definition; numbers in
roman refer to the \if@pageindex pages\else code lines \fi where
the entry is used.
\if@pageindex\else
\ifx\HLPrefix\@empty
The numbers preceded with `p.' are page numbers.
\else The numbers with no prefix are page numbers.
\fi\fi
\ifx\IndexLinksBlack\relax\else
All the numbers are hyperlinks.
% ^^A, they are made black just to let Adobe Reader work
% ^^A~faster.
\fi
\gmd@dip@hook% this hook is intended to let a~user add
% something without redefining the entire prologue, see below.
}}{}
% During the preparation of this package for publishing I~needed
% only to add something at the end of the default index prologue. So
% \DefIndex\gmd@dip@hook
\@emptify\gmd@dip@hook
\long\def\AtDIPrologue#1{\g@addto@macro\gmd@dip@hook{#1}}
% \changes{v0.98c}{06/9/8}{added because of need}
% The Author(s) of \docfm\ assume \pk{multicol} is known not to
% everybody. My assumption is the other so
\RequirePackage{multicol}
% \begin{quotation}If \pk{multicol} is in use, when the index is
% started we compute
% the remaining space on the current page; if it is greater than
% |\IndexMin|, the first part of the index will then be placed in the
% available space. The number of columns set is controlled by the
% counter |\c@IndexColumns| which can be changed with a
% |\setcounter| declaration.\end{quotation}
\newdimen\IndexMin \IndexMin = 133pt\relax% originally it was set
% 80\,pt, but with my default prologue there's at least 4.7\,cm needed
% to place the prologue and some index entries on the same page.
\newcount\c@IndexColumns \c@IndexColumns = 3
\renewenvironment{theindex}
{\begin{multicols}\c@IndexColumns[\index@prologue][\IndexMin]%
\IndexLinksBlack
\IndexParms \let\item\@idxitem \ignorespaces}%
{\end{multicols}}
\def\IndexLinksBlack{\hypersetup{linkcolor=black}}% To make Adobe
% Reader work faster.
\@ifundefined{IndexParms}
{\def\IndexParms{%
% \label{IndexParms}
\parindent \z@
\columnsep 15pt
\parskip 0pt plus 1pt
\rightskip 15pt
\mathsurround \z@
\parfillskip=-15pt plus 1 fil % \docfm\ defines this parameter
% rigid but that's because of the stretchable space (more
% precisely, a~|\dotfill|) between the item and the entries. But
% in \pk{gmdoc} we define no such special delimiters, so we add
% an ifinite stretch.
\small
\def\@idxitem{\par\hangindent 30pt}%
\def\subitem{\@idxitem\hspace*{15pt}}%
\def\subsubitem{\@idxitem\hspace*{25pt}}%
\def\indexspace{\par\vspace{10pt plus 2pt minus 3pt}}%
\ifx\EntryPrefix\@empty\else\raggedright\fi% long (actually,
% a~quite \emph{short but nonempty} entry prefix) made space
% stretches so terribly large in the justified paragraphs that
% we should make |\raggedright| rather.
\ifnum\c@IndexColumns>\tw@\raggedright\fi% the numbers in narrow
% col\-umns look better when they are |\raggedright| in my opinion.
}}{}
\def\PrintIndex{% we ensure the standard meaning of the line end
% character not to cause a~disaster.
\@ifQueerEOL{\StraightEOL\printindex\QueerEOL}%
{\printindex}}
% Remember that if you want to change not all the parameters, you
% don't have to redefine the entire |\IndexParms| macro but you may
% use a~very nice \LaTeX\ command |\g@addto@macro| (it has |\global|
% effect, also with an apeless name (|\gaddtomacro|) provided by
% \pk{gmutils}. (It adds its second argument at the end of definition
% of its first argument provided the first argument is a~no-argument
% macro.) Moreover, \pk{gmutils} provides also
% |\addtomacro| that has the same effect except it's not |\global|.
%
%
%
% \subdivision{The \ds\ directives}
%
%^^A \StraightEOL
%^^A \iffalse ^^A~ this passage is obsoleted
%^^A
%^^A This passage is obsoleted on 2006/11/30.
%^^A
%^^A In the \gmdoc\ thinking, the \ds\ directives belong to the
%^^A narration layer since they all begin with |%|. For now I~don't
%^^A have a~more minimal idea than to define a~pair of macros the first of
%^^A which would |\active|ate the |<| and the other, assigned to the
%^^A active |<|, would do the work and re`other' the |<| back.
%^^A
%^^A We provide two versions of the seargant-activating macro: the first
%^^A makes only the first subsequent seargant typeset a~\ds\
%^^A directive (and the previous meaning of the |<| will be restored, the
%^^A latter doesn't restore the previous meaning of |<| after (by) the
%^^A first occurrence and the latter obeys usual scoping rules while the
%^^A scoping rules of the first (|\docstrip|) are more strict: the scope
%^^A is delimited by the first occurrence of the subject~|<|.
%^^A
%^^A Let us remeber that, just as all the control sequences except
%^^A |\relax|, you can use |\docstrips| as an environment, i.e., in the
%^^A |\begin{docstrips}|\dots |\end{docstrips}| syntax. And, that
%^^A \env{macrocode} environment declares it.
%^^A \bgroup\catcode`\<=\active
%^^A \firstofone{\egroup
%^^A % \stanza
%^^A \newcommand*\docstrip{%
%^^A \gmd@storesearg
%^^A \catcode`\<=\active
%^^A \gmd@setdocstrips
%^^A \let\gmd@docstripshook=\gmd@restoresearg}%
%^^A % \stanza
%^^A \newcommand*\docstrips{%
%^^A \catcode`\<=\active
%^^A \gmd@setdocstrips
%^^A \@emptify\gmd@docstripshook}%
%^^A % \stanza
%^^A \def\gmd@setdocstrips{%
%^^A \def<{\ifmmode\sgtleftxii\else\afterfi\gmd@docstripdirective\fi}}
%^^A %\stanza
%^^A \def\gmd@storesearg{%
%^^A \edef\gmd@SeargantCat{\the\catcode`\<}%
%^^A \ifnum\gmd@SeargantCat=\active
%^^A \let\gmd@SeargantMng=<%
%^^A \fi}
%^^A % \stanza
%^^A \def\gmd@restoresearg{%
%^^A \catcode`\<=\gmd@SeargantCat\relax
%^^A \ifnum\gmd@SeargantCat=\active
%^^A \let<=\gmd@SeargantMng
%^^A \fi}%
%^^A % \stanza
%^^A }% of |\active| |<|'s |\firstofone|
%^^A
%^^A \fi ^^A~of the obsoletion iffalse
%^^A \QueerEOL
\foone{\@makeother\<\@makeother\>
\glet\sgtleftxii=<}
{
\def\gmd@docstripdirective{%
\begingroup\let\do=\@makeother
\do\*\do\/\do\+\do\-\do\,\do\&\do\|\do\!\do\(\do\)\do\>\do\<% ^^Ayes, we
% ^^A make \inverb|<|\catother\ since a~directive |<<|\<any text till ^^B>>
% ^^A~the end of line> theoretically may occur.
\@ifnextchar{<}{%
\let\do=\@makeother \dospecials
\gmd@docstripverb}
{\gmd@docstripinner}}%
\def\gmd@docstripinner#1>{%
\endgroup
\def\gmd@modulehashone{%
\Module{#1}\space
\@afternarrgfalse\@aftercodegtrue\@codeskipputgfalse}%
% ^^A \gmd@docstripshook
\gmd@textEOL\gmd@modulehashone}
% A~word of explanation: first of all, we
% close the group for changed |\catcode|s; the directive's text has its
% |\catcode|s fixed. Then we put the directive's text wrapped with the
% formatting macro into one macro in order to give just one token
% the \gmdoc's \TeX\ code scanner.
% ^^A But before we launch the \TeX\ code
% ^^A scanning with all the b\&w, we virtually restore the meaning of
% ^^A further |<|s to let them start further \ds\ directives,
% ^^A and after this possible restore (if it \emph{actually} takes place
% ^^A depends on the declaration used: |\docstrip| or |\docstrips|) we
% Then launch this big \TeX\ code scanning machinery by calling
% |\gmd@textEOL| which is an alias for the `narrative' meaning of the
% line end. This macro opens the verbatim group and launches the
% char-by-char scanner. That is this scanner because of what we
% encapsulated the directive's text with the formatting into one
% macro: to let it pass the scanner.
%
% That's why in the `old' macrocodes case the active |%| closes the
% group before launching |\gmd@docstripdirective|.
%
% \stanza
% The `verbatim' directive macro works very similarly.
}
\foone{\@makeother\<\@makeother\>
\glet\sgtleftxii=<
\catcode`\^^M=\active}%
{
\def\gmd@docstripverb<#1^^M{%
\endgroup%
\def\gmd@modulehashone{%
\ModuleVerb{#1}\@afternarrgfalse\@aftercodegtrue%
\@codeskipputgfalse}%
\gmd@docstripshook%
\gmd@textEOL\gmd@modulehashone^^M}%
}
% (\*Verbatim ;-) from \docfm:)
\providecommand*\Module[1]{{\mod@math@codes$\langle\mathsf{#1}\rangle$}}
\providecommand*\ModuleVerb[1]{{\mod@math@codes$\langle\langle\mathsf{#1}$}}
\def\mod@math@codes{\mathcode`\|="226A \mathcode`\&="2026 }
% \subdivision{The changes history}
%
% The contents of this section was copied \*verbatim from the
% \docfm's documentation, with only smallest necessary changes. Then ^^A((
% my additions were added :-))\,.
%
% \begin{quotation}To provide a change history log, the |\changes|
% command has been
% introduced. This takes [one optional and] three [mandatory]
% arguments, respectively, [the macro that'll become the entry's
% second level,] the version
% number of the file, the date of the change, and some detail
% regarding what change has been made [i.e., the description of the
% change]. The [second] of these arguments
% is otherwise ignored, but the others are written out and may be
% used to generate a history of changes, to be printed at the end of
% the document. [\dots\ I~ommit an obsolete remark about then-older
% MakeIndex's versions.]
%
% The output of the |\changes| command goes into the
% \<Glossary_File> and therefore uses the normal |\glossaryentry|
% commands. Thus MakeIndex or a~similar program can be used to
% process the output into a sorted ``glossary''. The |\changes|
% command commences by taking the usual measures to hide its
% spacing, and then redefines |\protect| for use within the argument
% of the generated |\indexentry| command. We re-code nearly all
% chars found in |\@sanitize| to letter since the use of special
% package which make some characters active might upset the
% |\changes| command when writing its entries to the file. However
% we have to leave |%| as
% comment and | | as \<space> otherwise chaos will happen. And, of
% course the |\| should be available as escape character.\end{quotation}
%
% We put the definition inside a~macro that will be
% executed by (the first use of) |\RecordChanges|. And we provide
% the default definition of |\changes| as a~macro just gobbling its
% arguments. We do this to provide no changes' writing out if
% |\RecordChanges| is not used.
%
%
% \changes[\changes]{v0.98e}{06/09/23}{definition put into a~macro that shall
% launch it when \cs{RecordChanges} is executed and the default
% defining it as a~gobbling macro}
%
\def\gmd@DefineChanges{%
\outer\long\def\changes{\@bsphack\begingroup\@sanitize
\catcode`\\\z@ \catcode`\ 10 \MakePercentIgnore
\MakePrivateLetters \StraightEOL
\MakeGlossaryControls
\changes@}}
\newcommand\changes[4][]{\PackageWarningNoLine{gmdoc}{%
^^JThe \bslash changes command used \on@line
^^Jwith no \string\RecordChanges\space declared.
^^JI shall not warn you again about it}%
%^^A\gmd@countlines{#4}%
\renewcommand\changes[4][]{%^^A\gmd@countlines{#4}
}}
\def\MakeGlossaryControls{%
\edef\actualchar{\string=}\edef\quotechar{\string!}%
\edef\levelchar{\string>}\edef\encapchar{\xiiclub}}%for the glossary the
% `actual', the `quote' and the `level' chars are respectively |=|,
% |!| and |>|, the `encap' char remains untouched. I~decided to
% preserve the \docfm's settings for the compatibility.
\newcommand\changes@[4][\generalname]{%
% \changes[\changes]{v0.97}{06/09/04}{The optional pseudo-argument
% added intended to be the macro-entry.}
\if@RecentChange{#3}% if the date is later than the one stored in
% \cs{c@Chang\+es\+Start\+Date},
\@tempswafalse
\ifx\generalname#1% then we check whether a~\CS-entry is given
% in the optional first argument or is it unchanged.
\ifx\last@defmark\relax\else% if no particular \CS is
% specified in |#1|, we check whether |\last@defmark| contains
% something and if so, we put it into |\gmu@tempb| scratch macro.
\@tempswatrue
\edef\gmu@tempb{% it's a~bug fix: while typesetting traditional
% \file{.dtx}es, \inverb|\last@defmark| came out with \inverb|\| at the
% beginning (which resulted with \inverb|\\|\<name> in the change
% log) but while typesetting the `new' way, it occurred
% without the bslash. So we gobble the bslash if it's
% present and two lines below we handle the exception of
% \inverb|\last@defmark|\equals|{\}| (what would happen if
% a~definition of \inverb|\\| was marked in new way \gmdoc ing).%^^A]
\if\bslash\last@defmark\else\last@defmark\fi}%
\ifx\last@defmark\bslash\let\gmu@tempb\last@defmark\fi%
\n@melet{gmd@glossCStest}{gmd/isaCS/\last@defmark}%
\fi
\else%the first argument isx not |\generalname| i.e.,
%a~particular \CS is specified by it (if some day one
% wishes to |\changes| |\generalname|, \heshe\ should type
% |\changes[generalname]|\dots)
\@tempswatrue
{\escapechar\m@ne
\xdef\gmu@tempb{\string#1}}%
\if\bslash\@xa\@firstofmany\string#1\relax\@@nil% we check
% whether \inverb|#1| is a~\CS\dots\ilrr
\def\gmd@glossCStest{1}%\dots\ and tell the glossary if so.
\fi
% ^^A~\@xa\quote@mname\@tempb\relax
\fi
\@ifundefined{gmd@glossCStest}{\def\gmd@glossCStest{0}}{}%
\protected@edef\gmu@tempa{\@nx\gmd@glossary{%
\if\relax\GeneralName\relax\else
\GeneralName% it's for the |\DocInclude| case to precede
% every |\changes| of the same file with the file name, cf.\ line
% \ref{GeneralName}.
\fi
#2\levelchar%
\if@tempswa% If the macro |\last@defmark|
% doesn't contain any \CS name (i.e., is empty) nor |#1|
% specifies a~\CS, the current
% changes entry was done at top-level. In this case we precede
% it by |\generalname|.
\gmu@tempb
\actualchar\bslash verb*%
\if\verbatimchar\gmu@tempb$\else\verbatimchar\fi
\if1\gmd@glossCStest\quotechar\bslash\fi \gmu@tempb
\if\verbatimchar\gmu@tempb$\else\verbatimchar\fi
\else
\space\actualchar\generalname
\fi
:\levelchar%^^A\gmd@deeolize{
#4%^^A}
}}%
\gmu@tempa
\grelaxen\gmd@glossCStest
\fi% of |\if@recentchange|
%^^A\gmd@countlines{#4}%
\endgroup\@esphack}
% Let's initialize |\last@defmark| and |\GeneralName|.
\@relaxen\last@defmark
\@emptify\GeneralName
\def\ChangesGeneral{\grelaxen\last@defmark}% If automatic detection of
% definitions is on, the default entry of |\changes| is the meaning of
% |\last@defmark|, the last detected definiendum that is. The
% declaration defined here serves to start a~scope of `general'
% \cs{changes}' entries.
\AtBegInput{\ChangesGeneral}
% Let's explain |\if@RecentChange|. We wish to check whether the
% change's date is later than date declared (if any limit date
% \emph{was} declared). First of all, let's establish a~counter to store
% the declared date. The untouched counters are equal 0 so if no date
% is declared there'll be no problem. The date will have the
% \<YYYYMMDD> shape both to be easily compared and readable.
\newcount\c@ChangesStartDate
\def\if@RecentChange#1{%
\gmd@setChDate#1\@@nil\@tempcnta
\ifnum\@tempcnta>\c@ChangesStartDate}
\def\gmd@setChDate#1/#2/#3\@@nil#4{% the last parameter will be a~|\count|
% register.
#4=#1\relax
\multiply#4 by\@M
\count8=#2\relax% I~know it's a~bit messy not to check whether the
% |#4| |\count| is |\count8| but I~know this macro will only be used
% with |\count0 | \nlpercent(\cs{@\+te\+m\+p\+cn\+ta}) and some
% higher (not a~scratch) one.
\multiply\count8 by100 %
\advance#4 by\count8 \count8=\z@
\advance#4 by#3\relax}
% Having the test defined, let's define the command setting the date
% counter. |#1| is to be the version and |#2| the date
% \cs[]{\{\<year>/\<month>/\<day>\}}.
\def\ChangesStart#1#2{%
% \changes{v0.98e}{06/09/23}{\cs{string}s added before the seargants
% in the message with account of \cs{docstrip(s)}}
\gmd@setChDate#2\@@nil\c@ChangesStartDate
\typeout{^^JPackage gmdoc info: ^^JChanges' start date #1 memorized
as \string<\the\c@ChangesStartDate\string> \on@line.^^J}
\advance\c@ChangesStartDate\m@ne% we shall show the changes \emph{at ^^B
% the specified day} and later.
\ifnum\c@ChangesStartDate>19820900 %\unskip\StraightEOL\footnote{^^A
% DEK writes in \textit{\TeX, The Program} of September 1982 as the
% date of \TeX\ Version 0.}\QueerEOL\label{TeXVer0}
% see \gmiflink[personalchanges]{below}.
\edef\gmu@tempa{%
\@nx\g@addto@macro\@nx\glossary@prologue{%
The changes
\if\relax\GeneralName\relax\else of \GeneralName\space\fi
earlier than
#1 \if\relax#1\relax #2\else(#2)\fi\space are not shown.}}%
\gmu@tempa
\fi}
% (Explanation to line \ref{TeXVer0}.)
% \gmhypertarget[personalchanges]{My} \TeX\ Guru has remarked that the
% change history tool should be used for documenting the changes that
% may be significant for the users not only for the author and talking
% of what may be significant to the user, no changes should be hidden
% since the first published version. However, the changes' start date
% may be used to provide hiding the author's `personal' notes: \heshe\
% should only date the `public' changes with the four digit year and
% the `personal' ones with two digit year and set
% |\ChangesStart{}{1000/0/0}| or so.
%
% In line \ref{TeXVer0} I~establish a~test value that corresponds to
% a~date earlier than any \TeX\ stuff and is not too small (early) to
% ensure that hiding the two digit year changes shall not be mentioned
% in the changes prologue.
%
% \changes{v0.98d}{2006/9/15}{An entry to show the change history
% works: watch and admire. Some sixty \cs{changes} entries irrelevant
% for the users-other-than-myself are hidden due to the trick
% described on
% p.\protect\,\protect\pageref{personalchanges}.%
% \protect\StoreMacro\protect\hyperpage
% \protect\def\protect\hyperpage#1{\protect\RestoreMacro\protect\hyperpage}%
% \protect\gobble}
% \stanza
% \begin{quotation}The entries [of a~given version number] are sorted
% for convenience by the name of [the macro explicitly specified as
% the first argument or] the most recently introduced macroname
% (i.e., that in the most recent |\begin{macro}| command [or
% |\Define|]). We therefore provide [|\last@defmark|] to
% record that argument, and provide a default definition in case
% |\changes| is used outside a \env{macro} environment. (This is a
% wicked hack to get such entries at the beginning of the sorted
% list! It works providing no macro names start with |!| or |"|.)
%
% This macro holds the string placed before changes entries on
% top-level.\end{quotation}
\def\generalname{General}
% \changes[\chschange]{v0.98a}{06/09/06}{added.}
%
% \stanza
% \begin{quotation}To cause the changes to be written (to a \pk{.glo})
% file, we define |\RecordChanges|
% to invoke \LaTeX's usual |\makeglossary| command.\end{quotation}
%
% I~add to it also the |\write|ing definition of the |\changes| macro
% to ensure no changes are written out without |\RecordChanges|.
%
\def\RecordChanges{\makeglossary\gmd@DefineChanges
\@relaxen\RecordChanges}
% \changes{v0.98c}{06/09/08}{made self-\cs{relax}ing}
% \changes{v0.98e}{06/09/23}{made it defining \cs{changes}}
% \begin{quotation}The remaining macros are all analogues of those used
% for the \env{theindex} environment.
% When the glossary is started we compute the space which remains at
% the bottom of the current page; if this is greater than |\GlossaryMin| then the
% first part of the glossary will be placed in the available space. The number of
% columns set [is] controlled by the counter |\c@GlossaryColumns| which can be
% changed with a |\setcounter| declaration.\end{quotation}
% \begin{CodeSpacesBlank}
\newdimen\GlossaryMin \GlossaryMin = 80pt
% \DefIndex*\c@GlossaryColumns
\newcount\c@GlossaryColumns \c@GlossaryColumns = 2
% \end{CodeSpacesBlank}
%\begin{quotation}The environment \env{theglossary} is defined in the
%same manner as the \env{theindex}
% environment.\end{quotation}
%
\newenvironment{theglossary}{%
% \changes{v0.99m}{2008/08/10}{added \cs{IndexLinksBlack}}
\begin{multicols}\c@GlossaryColumns
[\glossary@prologue][\GlossaryMin]%
\GlossaryParms \IndexLinksBlack
\let\item\@idxitem \ignorespaces}%
{\end{multicols}}
% Here is the MakeIndex style definition:
% \index{gmglo.ist@\file{gmglo.ist}}
%^^A \begin{docstrips}
%</package>
%<+gmglo>preamble
%<+gmglo>"\n \\begin{theglossary} \n
%<+gmglo>\\makeatletter\n"
%<+gmglo>postamble
%<+gmglo>"\n\n \\end{theglossary}\n"
%<+gmglo>keyword "\\glossaryentry"
%<+gmglo>actual '='
%<+gmglo>quote '!'
%<+gmglo>level '>'
%<*package>
%^^A \end{docstrips}
% The MakeIndex shell command for the glossary should look as
% follows:
% \[|makeindex -r -s gmglo.ist -o |\<myfile>|.gls |\<myfile>|.glo|\]
% ^^B
% ^^B makeindex -r -s gmglo.ist -o <myfile>.gls <myfile>.glo
% ^^B
% where |-r| commands MakeIndex not to make implicit page ranges,
% |-s| commands MakeIndex to use the style stated next not the
% default settings and the |-o| option with the subsequent filename
% defines the name of the output.
%
% \begin{quotation}The |\GlossaryPrologue| macro is used to place
% a~short message above the glossary into the document. It is
% implemented by redefining |\glossary@prologue|, a macro which
% holds the default text. We better make it a long macro to allow
% |\par| commands in its argument.\end{quotation}
%
\long\def\GlossaryPrologue#1{\@bsphack
\def\glossary@prologue{#1}%
\@esphack}
% \begin{quotation}Now we test whether the default is already defined
% by another package file. If not we define it.\end{quotation}
% \changes{v0.98a}{06/09/06}{a~bug fixed: \cs{filediv} replaced with
% \cs{indexdiv} in the default prologue.}
\@ifundefined{glossary@prologue}
{\def\glossary@prologue{\indexdiv{{Change History}}%
\markboth{{Change History}}{{Change History}}%
}}{}
%\begin{quotation}Unless the user specifies otherwise, we set the
%change history using the same parameters as for the index.\end{quotation}
% \Define\GlossaryParms
\AtBeginDocument{%
\@ifundefined{GlossaryParms}{\let\GlossaryParms\IndexParms}{}}
%\begin{quotation}
% To read in and print the sorted change history, just put the
% |\PrintChanges| command as the last (commented-out, and thus
% executed during the documentation pass through the file) command in
% your package file. Alternatively, this command may form one of the
% arguments of the |\StopEventually| command, although a change history
% is probably not required if only the description is being printed.
% The command assumes that MakeIndex or some other program has
% processed the \pk{.glo} file to generate a sorted \pk{.gls}
% file.\end{quotation} \Define\PrintChanges
\def\PrintChanges{% to avoid a~disaster among queer EOLs:
\@ifQueerEOL
{\StraightEOL\@input@{\jobname.gls}\QueerEOL}%
{\@input@{\jobname.gls}}%
\g@emptify\PrintChanges}
\pdef\toCTAN#1#2{%\
% \begin{enumargs}
% \item version number,
% \item date \<year/month/day>.
% \end{enumargs}
% \changes{v0.99r}{2008/11/22}{added}
\changes{#1}{#2}{put to \acro{CTAN} on #2}}
%
% \subdivision{The checksum}
%
%
% \docfm\ provides a~checksum mechanism that counts the backslashes
% in the scanned code. Let's do almost the same.
%
% At the beginning of the source file you may put the |\CheckSum|
% macro with a~number (in one of \TeX's formats) as its argument and
% \TeX\ with \pk{gmdoc} shall count the number of the \emph{escape ^^B
% chars} in the source file and tell you in the \file{.log} file (and
% on the terminal)
% whether you have typed the right number. If you don't type |\CheckSum|,
% \TeX\ anyway will tell you how much it is.
\newcount\check@sum
\def\CheckSum#1{\@bsphack\global\check@sum#1\relax\@esphack}
\newcounter{CheckSum}
% \label{checksum}
\newcommand*\step@checksum{\stepcounter{CheckSum}}
% And we'll use it in the line \ref{checksumUse} (|\stepcounter| is
% |\global|). See also the |\chschange| declaration,
% l.\,\ref{chschange}.
%
% However, the check sum mechanism in \gmdoc\ behaves slightly
% different than in \docfm\ which
% is nicely visible while \pk{gmdoc}ing \docfm: \docfm\ states its
% check sum to be 2171 and our count counts 2126. The mystery lies in
% the fact that \docfm's CheckSum mechanism counts the code's
% backslashes no matter what they mean and the \gmdoc's the escape
% chars so, among others, |\\| at the default settings increases \docfm's
% CheckSum by 2 while the \gmdoc's by 1. (There are 38 occurrences of |\\|
% in \pk{doc.dtx} \env{macrocode}s, I~counted
% myself.)\begin{StraightEOL}\footnote{My opinion is that nowadays
% a~check sum
% is not necessary for checking the completness of a~file but
% I~like it as a~marker of file development and this more than
% that is its r\ocircum le in \gmdoc.}
% \end{StraightEOL}
%
%
% \begin{quotation}But |\Finale| will be called at the very end of a
% file. This is exactly the point were we want to know if the file
% is uncorrupted. Therefore we also call |\check@checksum| at this
% point.\end{quotation}
% In \pk{gmdoc} we have the |\AtEndInput| hook.
\AtEndInput{\check@checksum}
% Based on the lines 723--741 of \pk{doc.dtx}.
\def\check@checksum{\relax
\ifnum\check@sum=\z@
\edef\gmu@tempa{% why \cs{edef}---see line \ref{wypis sumy}
\@nx\typeout{**********************************^^J%
* The input file \gmd@inputname\space has no Checksum
stated.^^J%
* The current checksum is \the\c@CheckSum.^^J%
\gmd@chschangeline% a~check sum changes history entry, see below.
* (package gmdoc info.)^^J%
**********************************^^J}}
\else
\ifnum\check@sum=\c@CheckSum
\edef\gmu@tempa{%
\@nx\typeout{*****+*+*+*+*+*+*+*+*+*+^^J%
* The input file \gmd@inputname: Checksum passed.^^J%
\gmd@chschangeline
* (package gmdoc info.)^^J%
*****+*+*+*+*+*+*+*+*+*+^^J}}
\else
\edef\gmu@tempa{%
\@nx\typeout{********!*!*!*!*!*!*!*!*!*!*!*!^^J%
*! The input file \gmd@inputname:^^J%
*! The CheckSum stated: \the\check@sum\space<> my
count: \the\c@CheckSum.^^J%
\gmd@chschangeline
*! (package gmdoc info.)^^J%
********!*!*!*!*!*!*!*!*!*!*!*!^^J}}%
\fi
\fi
\gmu@tempa
\@xa\AtEndDocument\@xa{\gmu@tempa}%\label{wypis sumy} we print
% the checksum notification on the terminal immediately and at end
% of \TeX ing not to have to scroll the output far nor search the log.
\global\check@sum\z@}
% As I~mentioned above, I~use the check sum mechanism to mark the file
% growth. Therefore I~provide a~macro that produces a~line on the
% terminal to be put somewhere at the
% beginning of the source file's commentary for instance.
\def\gmd@chschangeline{%
\xiipercent\space\string\chschange
{\@ifundefined{fileversion}{v???}{\fileversion}}%
{\the\year/\the\month/\the\day}%
{\the\c@CheckSum}^^J%
\xiipercent\space\string\chschange
{\@ifundefined{fileversion}{v???}{\fileversion}}%
{\@xa\@gobbletwo\the\year/\the\month/\the\day}%
{% with two di\-g\-it year in case you use |\ChangesStart|.
\the\c@CheckSum}^^J}
% And here the meaning of such a~line is defined:
\newcommand*\chschange[3]{%\label{chschange}
\csname changes\endcsname{#1}{#2}{CheckSum #3}% |\csname...| because
% \nlpercent\cs{cha\+n\+g\+es} is \inverb|\outer|.
\CheckSum{#3}}
% It will make a~`General' entry in the change history unless
% used in some |\Define|'s scope or inside a~\env{macro}
% environment. It's intended to be put
% somewhere at the beginning of the documented file.
%
%
% \changes[\CheckSum]{v0.98a}{06/09/06}{`package gmdoc info'
% statement moved to the end of the checksum messages.}
% \changes[\CheckSum]{v0.98b}{06/09/07}{typing out the
% \cs{chschange} line added}
%
%
% \subdivision{Macros from \pk{ltxdoc}}
%
% I'm not sure whether this package still remains `minimal' but
% I~liked the macros provided by \pk{ltxdoc.cls} so much\dots
% \stanza
%
% The next page setup declaration is intended to be used
% with the \pk{article}'s default Letter paper size. But since
\newcommand*\ltxPageLayout{%
%\begin{quotation}Increase the text width slightly so that width the standard fonts
% 72 columns of code may appear in a |macrocode| environment.\end{quotation}
%
\setlength{\textwidth}{355pt}%
% \begin{quotation}Increase the marginpar width slightly, for long
% command names. And increase the left margin by a similar
% amount.\end{quotation}
% To make these settings independent from the defaults (changed e.g.\
% in \pk{gmdocc.cls}) we replace the original |\addtolength|s with
% |\setlength|s.
%^^A \addtolength\marginparwidth{30pt}%
%^^A \addtolength\oddsidemargin{20pt}%
%^^A \addtolength\evensidemargin{20pt}
\setlength\marginparwidth{95pt}%
\setlength\oddsidemargin{82pt}%
\setlength\evensidemargin{82pt}}
% %\skiplines
% \def\task#1#2{}
% % What is this for? Probably to turn an obsolete command off. Or maybe
% % a~\acro{TODO}?
% % \endskiplines
%\subdivision{\cs{DocInclude} and the \pk{ltxdoc}-like setup}
%
% Let's provide a~command for including
% multiple files into one document. In the \pk{ltxdoc} class such
% a~command is defined to include files as parts. But we prefer to
% include them as chapters in the classes that provide
% |\chapter|. We'll redefine |\maketitle| so that it make a~chapter or
% a~part heading \emph{unlike} in \pk{ltxdoc} where the file parts
% have their titlepages with only the filename and \pk{article}-like titles
% made by |\maketitle|.
%
% But we will also provide a~possibility of typesetting multiple files
% exactly like with the \pk{ltxdoc} class.
%
% \begin{macro}{\DocInclude}
%
% So, define the |\DocInclude| command, that acts
% \begin{quotation}more or less exactly the same as |\include|, but
% uses |\DocInput| on a \pk{dtx} [or \pk{.fdd}] file, not |\input| on
% a \pk{tex} file.\end{quotation}
% Our version will accept also \pk{.sty}, \pk{.cls}, and \pk{.tex}
% files.
%
% \changes{v0.98}{06/09/05}{\cs{@makeother\protect\bslash_} added}
\newcommand*\DocInclude{\bgroup\@makeother\_\Doc@Include}% First, we
% make \inverb|_| `other' in order to allow it in the filenames.
\newcommand*{\Doc@Include}[2][]{% originally it took just one
% argument. Here we make it take two, first of which is intended to
% be the path (with the closing \inverb |/|). This is intended not to print
% the path in the page footers only the filename.\par
% \CodeUsage\HLPrefix
\egroup% having the arguments read, we close the group opened by the
% previous macro for |_|\catother.
\gdef\HLPrefix{\filesep}%
\gdef\EntryPrefix{\filesep}% we define two rather kernel parameters
% to expand to the file marker. The first will bring the information
% to one of the default \inverb|\IndexPrologue|'s
% |\if|s. Therefore the
% definition is global. The latter is such for symmetry.
\def\GeneralName{#2\actualchar\pk{#2} }% \label{GeneralName} for the
% changes'history main level entry.
%
% Now we check whether we try to include ourselves and if
% so---we'll (create and) read an \file{.auxx} file instead of
% (the main) \file{.aux} to avoid an infinite recursion of |\input|s.
%
\edef\gmd@jobname{\jobname}%
\edef\gmd@difilename{% we want the filename all `other', just as
% in \inverb|\jobname|.
\@xa\@xa\@xa\@gobble\@xa\string\csname#2\endcsname}%
\ifx\gmd@jobname\gmd@difilename
\def\gmd@auxext{auxx}%
\else
\def\gmd@auxext{aux}%
\fi
\relax
%^^A\if@ltxDocInclude
\clearpage
% ^^A\fi
\gmd@docincludeaux
\def\currentfile{gmdoc-IncludeFileNotFound.000}%
\let\fullcurrentfile\currentfile
\IfFileExists{#1#2.fdd}{\edef\currentfile{#2.fdd}}{% it's not \pk{.fdd},
\IfFileExists{#1#2.dtx}{\edef\currentfile{#2.dtx}}{% it's not \pk{.dtx}
% either,
\IfFileExists{#1#2.sty}{\edef\currentfile{#2.sty}}{% it's not
% \pk{.sty},
\IfFileExists{#1#2.cls}{\edef\currentfile{#2.cls}}{% it's
% not \pk{.cls},
\IfFileExists{#1#2.tex}{\edef\currentfile{#2.tex}}{% it's
% not \pk{.tex},
\IfFileExists{#1#2.fd}{\edef\currentfile{#2.fd}}{% so it
% must be \pk{.fd} or error.
\PackageError{gmdoc}{\string\DocInclude\space file
#1#2.fdd/dtx/sty/cls/tex/fd not found.}}}}}}}%
% \changes{v0.98j}{06/10/16}{\pk{.fdd} added in \cs{DocInclude}'s
% search for the extension.}
\edef\fullcurrentfile{#1\currentfile}%
\ifnum\@auxout=\@partaux
\@latexerr{\string\DocInclude\space cannot be nested}\@eha
\else \@docinclude{#1}#2 \fi}% Why is |#2| delimited with | | not
% braced as we are used to, one may ask.
% \changes[\DocInclude]{v0.97}{06/09/03}{\cs{@docinclude} gets 2 params
% in order to \cs{includeonly} work as expected, with only the names
% (w.o. the paths)}
\def\@docinclude#1#2 {% To match the macro's parameter string, is an
% answer. But why is |\@docinclude| defined so? Originally, in
% \pk{ltxdoc} it takes one argument and it's delimited with a~space
% probably in resemblance to the true |\input| \nlpercent(|\@@input| in
% \LaTeX). ^^A\if@ltxDocInclude
\clearpage
%^^A\fi
\if@filesw \gmd@writemauxinpaux{#2.\gmd@auxext}\fi% this strange macro with
% a~long name is another thing to allow |_| in the filenames (see
% line \ref{gmd@writemauxinpaux}).
\@tempswatrue
\if@partsw \@tempswafalse\edef\gmu@tempb{#2}%
\@for \gmu@tempa:=\@partlist\do{\ifx\gmu@tempa\gmu@tempb\@tempswatrue\fi}%
\fi
\if@tempswa \let\@auxout\@partaux
\if@filesw
\immediate\openout\@partaux #2.\gmd@auxext\relax% Yes, only |#2|. It's to
% create and process the partial \pk{.aux(x)} files always in the main
% document's (driver's) directory.
% \changes[\DocInclude]{v0.97}{06/09/04}{\protect\TeX forced to
% process the partial \pk{.aux}es in the main document directory}
\immediate\write\@partaux{\relax}%
\fi
% \begin{quotation}We need to save (and later restore) various
% index-related commands which might be changed by the included
% file.\end{quotation}
% \changes{v0.98a}{06/09/06}{\cs{MacroStoringDo} renamed (in
% \pk{gmutils}) to \cs{StoringAndRelaxingDo}, and
% \cs{MacroRestoringDo} to \cs{RestoringDo}.}
\StoringAndRelaxingDo\gmd@doIndexRelated
\if@ltxDocInclude\part{\currentfile}% In the \pk{ltxdoc}-like setup
% we make a~part title page with only the filename and the file's
% |\maketitle| will typeset an \pk{article}-like title.
\else\let\maketitle=\InclMaketitle
\fi% In the default setup we
% redefine |\maketitle| to typeset a~common chapter or part heading.
\if@ltxDocInclude\xdef@filekey\fi
\GetFileInfo{\currentfile}% it's my (GM) addition with the account
% of using file info in the included files' title/\:heading etc.
\incl@DocInput{\fullcurrentfile}% originally just |\currentfile|.
\if@ltxDocInclude\else\xdef@filekey\fi% in the default case we add
% new file to the file key \emph{after} the input because in this
% case it's the files own |\maketitle| what launches the
% sectioning command that increases the counter.
% \par And here is the moment to restore the index-related
% commands.
\RestoringDo\gmd@doIndexRelated
% ^^A\if@ltxDocInclude
\clearpage
% ^^A\fi
\gmd@writeckpt{#1#2}%
\if@filesw \immediate\closeout\@partaux \fi
\else\@nameuse{cp@#1#2}%
\fi
\let\@auxout\@mainaux}% end of |\@docinclude|.
% \end{macro}
%
% (Two is a~sufficient number of iterations to define a~macro for.)
\def\xdef@filekey{{\@relaxen\ttfamily% \label{LetTTFRelax}This
% assignment is very trickly crafted: it makes \emph{all}
% |\ttfamily|s present in the |\filekey|'s expansion unexpandable
% not only the one added in this step.
\xdef\filekey{\filekey, \thefilediv={\ttfamily\currentfile}}}}
% To allow |_| in the filenames we must assure |_| will be \catother\
% while reading the filename. Therefore define
% \DefIndex\gmd@writemauxinpaux
\def\gmd@writemauxinpaux#1{% \label{gmd@writemauxinpaux}
% this name comes from `\emph{write} outto \emph main
% \emph{\pk{.aux}} to \emph{in}put \emph partial
% \emph{\pk{.aux}}'.\par
% We wrap |\@input{|\<partial \pk{.aux}>|}| in a~|_|\catother\
% hacked scope. This hack is especially recommended here since the
% \pk{.aux} file may contain a~non-|\global| stuff that should not
% be localized by a~group that we would have to establish if we
% didn't use the hack. (Hope you understand it. If not, notify me
% and for now I'll only give a~hint: ``Look at it with the \TeX's
% eyes''. More uses of this hack are to be seen in \pk{gmutils}
% where they are a~bit more explained.)
\immediate\write\@mainaux{%
\bgroup\string\@makeother\string\_%
\string\firstofone{\egroup
\string\@input{#1}}}}
% We also slightly modify a~\LaTeX\ kernel macro |\@writeckpt| to
% allow |_| in the file name.
% \changes[\DocInclude]{v0.98a}{06/09/06}{\cs{@writeckpt} wrapped
% with \cs{firstofone} hack to allow \cs[]{_} in the file names.}
% \DefIndex\gmd@writeckpt
\def\gmd@writeckpt#1{%
\immediate\write\@partaux{%
\string\bgroup\string\@makeother\string\_%
\string\firstofone\@charlb\string\egroup}
\@writeckpt{#1}%
\immediate\write\@partaux{\@charrb}}
\def\gmd@doIndexRelated{%
\do\tableofcontents \do\makeindex \do\EnableCrossrefs
\do\PrintIndex \do\printindex \do\RecordChanges \do\PrintChanges
\do\theglossary \do\endtheglossary}
\@emptify\filesep
% The \pk{ltxdoc} class establishes a~special number format for
% multiple file documentation numbering needed to document the \LaTeX\
% sources. I~like it too, so
\def\aalph#1{\@aalph{\csname c@#1\endcsname}}
\def\@aalph#1{%
\ifcase#1\or a\or b\or c\or d\or e\or f\or g\or h\or i\or
j\or k\or l\or m\or n\or o\or p\or q\or r\or s\or
t\or u\or v\or w\or x\or y\or z\or A\or B\or C\or
D\or E\or F\or G\or H\or I\or J\or K\or L\or M\or
N\or O\or P\or Q\or R\or S\or T\or U\or V\or W\or
X\or Y\or Z\else\@ctrerr\fi}
% A~macro that initialises things for |\DocInclude|.
\def\gmd@docincludeaux{%
% We set the things for including the files only once.
\global\@relaxen\gmd@docincludeaux
% By default, we will include multiple files into one document
% as chapters in the classes that provide |\chapter| and as parts
% elsewhere.
\ifx\filediv\relax
\ifx\filedivname\relax% (nor |\filediv| neither
% |\filedivname| is defined by the user)
% \changes[\filediv]{v0.98a}{06/09/06}{definition moved into
% \cs{DocInclude} and let \cs{relax} by default ($\quotechar=$a~bug fix).}
\@ifundefined{chapter}{%
\SetFileDiv{part}}%\changes[\DocInclude]{v0.98f}{06/9/29}{\cs{filediv(name)}
% defined as \cs[]{(\protect\bslash)part} for the classes that
% don't know \cs{chapter}}
{\SetFileDiv{chapter}}%
\else% (|\filedivname| is defined by the user, |\filediv| is not)
\SetFileDiv{\filedivname}% why not? Inside is |\edef| so it'll work.
\fi
\else% (|\filediv| is defined by the user
\ifx\filedivname\relax% and |\filedivname| is not)
%^^A~I~deleted a~six level test which one sectioning command
%^^A~|\filedivname| isx.
\PackageError{gmdoc}{You've redefined \string\filediv\space
without redefining \string\filedivname.}{Please redefine the
two macros accordingly. You may use \string\SetFileDiv{name
without bslash}.}%
\fi
\fi
% \changes[\DocInclude]{v0.99m}{2008/08/09}{
% resetting of codeline number with every
% \cs{filedivname} commented out because with the
% \env{countalllines} option it caused that reset at \cs{maketitle}
% after some lines of file}
% ^^A \@addtoreset{codelinenum}{\filedivname}% remember it has
% ^^A~a~|\global| effect in fact. For each file we'll reset
% ^^A~\env{codelinenum}.
\def\thefilediv{\aalph{\filedivname}}% The files will be numbered
% with letters, lowercase first.
\@xa\let\csname the\filedivname\endcsname=\thefilediv% This
% li\-ne lets |\the|\<chapter> etc.\ equal |\thefilediv|.
\def\filesep{\thefilediv-}% File separator (identifier) for the index.
\let\filekey=\@gobble
\g@addto@macro\index@prologue{%
\gdef\@oddfoot{\parbox{\textwidth}{\strut\footnotesize
\raggedright{\bfseries File Key:} \filekey}}% The footer for
% the pages of index.
\glet\@evenfoot\@oddfoot}% \label{oneside}anyway, it's intended to
% be oneside.
\g@addto@macro\glossary@prologue{%
\gdef\@oddfoot{\strut Change History\hfill\thepage}% The footer
% for the changes history.
\glet\@evenfoot\@oddfoot}%
% \changes[\DocInclude]{v0.98c}{06/9/10}{Change History footer
% defined}
\gdef\@oddfoot{% The footer of the file pages will be its name and,
% if there is a~file info, also the date and version.
\@xa\ifx\csname ver@\currentfile\endcsname\relax
File \thefilediv: {\ttfamily\currentfile} %
\else
\GetFileInfo{\currentfile}%
File \thefilediv: {\ttfamily\filename} %
Date: \filedate\ %
Version \fileversion
\fi
\hfill\thepage}%
\glet\@evenfoot\@oddfoot% see line \ref{oneside}.
% ^^A \show\filedivname% for debug
\@xa\def\csname\filedivname name\endcsname{File}% we
% redefine the name of the proper division to `File'.
\ifx\filediv\section
\let\division=\subsection
\let\subdivision=\subsubsection
\let\subsubdivision=\paragraph
% If |\filediv| is higher than
% |\section| we don't change the three divisions (they are
% |\section|, |\subsection| and |\subsubsection| by default).
% |\section| seems to me the lowest reasonable sectioning command
% for the file. If |\filediv| is lower you should rather rethink
% the level of a~file in your documentation not redefine the two
% divisions.
\fi}% end of |\gmd@docincludeaux|.
% The |\filediv| and |\filedivname| macros should always be set
% together. Therefore provide a~macro that takes care of both at
% once. Its |#1| should be a~sectioning name without the backslash.
\def\SetFileDiv#1{%
\edef\filedivname{#1}%
\@xa\let\@xa\filediv\csname#1\endcsname}
% \changes{v0.98c}{06/9/10}{added and a~bug fixed in \cs{docincludeaux}}
\def\SelfInclude{\DocInclude{\jobname}}
% \changes{v0.98a}{06/09/06}{added in response to the needs of the
% first user, of me that is}
% \changes{v0.99l}{2008/08/05}{Made a~shorthand for
% \cs{Docinclude}\cs{jobname} instead of repeating 99\% of
% \cs{DocInclude}'s code}
% The \pk{ltxdoc} class makes some preparations for inputting multiple
% files. We are not sure if the user wishes to use \pk{ltxdoc}-like
% way of documenting (maybe \heshe\ will prefer what I~offer,
% \pk{gmdocc.cls} e.g.), so we put
% those preparations into a~declaration.
%
\newif\if@ltxDocInclude
\newcommand*\ltxLookSetup{%
\SetFileDiv{part}%
\ltxPageLayout
\@ltxDocIncludetrue
}
\@onlypreamble\ltxLookSetup
% The default is that we |\DocInclude| the files due to the original
% \gmdoc\ input settings.
\let\incl@DocInput=\DocInput
\@emptify\currentfile% for the pages outside the
% |\DocInclude|'s scope. In force for all includes.
%
%^^A \iffalse
%^^A % \stanza
%^^A % But it may be not a~usual situation to include the source files
%^^A % as with \pk{ltxdoc} but with the \emph{new} |\DocInput| macro. So
%^^A % let's |\def|
%^^A % \Define\ltxMultidocOldSetup
%^^A \newcommand*\ltxMultidocOldSetup{%
%^^A \ltxMultidocSetup
%^^A \let\incl@DocInput=\OldDocInput}
%^^A
%^^A \@onlypreamble\ltxMultidocOldSetup
%^^A
%^^A And, for the symmetry, if you prefer the look offered by
%^^A me, but
%^^A \fi
% If you want to |\Doc/SelfInclude| \docfm-likes:
\newcommand*\olddocIncludes{%
\let\incl@DocInput=\OldDocInput}
% And, if you have set the previous and want to set it back:
\newcommand*\gmdocIncludes{%
\let\incl@DocInput=\DocInput
\AtBegInput{\QueerEOL}}% to move back the |\StraightEOL| declaration put
% at begin input by |\olddocIncludes|.
%
%
% \subdivision{Redefinition of \cs{maketitle}}
%
% A~not-so-slight alteration of the \TextUsage\maketitle\ command in
% order it allow multiple titles in one document seems to me very
% clever. So let's copy again (\pk{ltxdoc.dtx} the lines 643--656):
%
% \begin{quotation}The macro to generate titles is easily altered in
% order that it can be used more than once (an article with many
% titles). In the original, diverse macros were concealed
% after use with |\relax|. We must cancel anything that may have been
% put into |\@thanks|, etc., otherwise all titles will carry forward
% any earlier such setting!\end{quotation}
% But here in \pk{gmdoc} we'll do it locally for (each) input
% not to change the main title settings if there are any.
\AtBegInput{%^^A~ why provide not just \cs{def}?
\providecommand*\maketitle{\par
\begingroup \def \thefootnote {\fnsymbol {footnote}}%
\setcounter {footnote}\z@
\def\@makefnmark{\hbox to\z@{$\m@th^{\@thefnmark}$\hss}}%
\long\def\@makefntext##1{\parindent 1em\noindent
\hbox to1.8em{\hss$\m@th^{\@thefnmark}$}##1}%
\if@twocolumn \twocolumn [\@maketitle ]%
\else \newpage \global \@topnum \z@ \@maketitle \fi
% \begin{quotation}For special formatting requirements (such as in
% \acro{TUG}boat), we use pagestyle |titlepage| for this; this is later defined
% to be |plain|, unless already defined, as, for example, by
% \pk{ltugboat.sty}.\end{quotation}
\thispagestyle{titlepage}\@thanks \endgroup
%\begin{quotation}If the driver file
% documents many files, we don't want parts of a title of one to
% propagate to the next, so we have to cancel these:\end{quotation}
\setcounter {footnote}\z@
\gdef\@date{\today}\g@emptify\@thanks%
\g@emptify\@author\g@emptify\@title%
}%\par
% \begin{quotation}When a number of articles are concatenated into
% a~journal, for example, it is not usual for the title pages of
% such documents to be formatted differently. Therefore, a class
% such as \pk{ltugboat} can define this macro in advance. However,
% if no such definition exists, we use pagestyle plain for title
% pages.\end{quotation}
\@ifundefined{ps@titlepage}{\let\ps@titlepage=\ps@plain}{}%
% \par And let's provide |\@maketitle| just in case: an error occurred
% without it at \TeX ing with \pk{mwbk.cls} because this class with the
% default options does not define |\@maketitle|. The below definitions
% are taken from \pk{report.cls} and \pk{mwrep.cls}.
\providecommand*\@maketitle{%
\newpage\null \vskip 2em\relax%
\begin{center}%
\titlesetup
\let \footnote \thanks
{\LARGE \@title \par}%
\vskip 1.5em%
{\large \lineskip .5em%
\begin{tabular}[t]{c}%
\strut \@author
\end{tabular}\par}%
\vskip 1em%
{\large \@date}%
\end{center}%
\par \vskip 1.5em\relax}%\par
% We'd better restore the primary meanings of the macros making
% a~title. (\LaTeXe\ source, File F: ltsect.dtx Date: 1996/12/20 Version
% v1.0z, lines 3.5.7.9--12.14--17.)
\providecommand*\title[1]{\gdef\@title{#1}}
\providecommand*\author[1]{\gdef\@author{#1}}
\providecommand*\date[1]{\gdef\@date{#1}}
\providecommand*\thanks[1]{\footnotemark
\protected@xdef\@thanks{\@thanks
\protect\footnotetext[\the\c@footnote]{#1}}%
}%
\providecommand*\and{%| % \begin{tabular}|
\end{tabular}%
\hskip 1em \@plus.17fil%
\begin{tabular}[t]{c}}%| % \end{tabular}|
% And finally, let's initialize \cs{tit\+le\+set\+up} if it is not yet.
\providecommand*\titlesetup{}%
}% end of |\AtBegInput|.
%
% \changes[\titlesetup]{v0.98c}{06/09/08}{\cs{provide}d \cs[]{\{\}}}
%
%
% The \pk{ltxdoc} class redefines the |\maketitle| command to allow
% multiple titles in one document. We'll do the same and something
% more: our |\Doc/SelfInclude| will turn the file's |\maketitle| into
% a~part or chapter heading. But, if hte |\ltxLookSetup| declaration
% is in force, |\Doc/SelfInclude| will make for an included file
% a~part's title page and an \pk{article}-like title.
%
% \stanza
% Let's initialize the file division macros.
\@relaxen\filediv
\@relaxen\filedivname
\@relaxen\thefilediv
% \changes[\thefilediv]{v0.99m}{2008/08/06}{let to \cs{relax} by
% default}
% If we don't include files the \pk{ltxdoc}-like way, we wish to
% redefine |\maketitle| so that it typesets a~division's heading.
%
% Now, we redefine |\maketitle| and its relatives.
%
\def\InclMaketitle{%
% \changes{v0.98a}{06/09/05}{renamed from \cs{incl@maketitle} because
% I~needed it in self-input hacks.}
{\def\and{, }% we make |\and| just a~comma.
{\let\thanks=\@gobble% for the toc version of the heading we
% discard |\thanks|.
\protected@xdef\incl@titletotoc{\@title\if@fshda\protect\space
(\@author)\fi}% we add the author iff the `files have
% different authors' \nlpercent(|@fshda|)
}%
\def\thanks##1{\footnotemark
\protected@xdef\@thanks{\@thanks% to keep the previous |\thanks|
% if there were any.
\protect\footnotetext[\the\c@footnote]{##1}}}% for some
% mysterious reasons so defined |\thanks| do typeset the footnote
% mark and text but they don't hyperlink it
% properly. A~\pk{hyperref} bug?
\@emptify\@thanks
\protected@xdef\incl@filedivtitle{%
[{\incl@titletotoc}]% braces to allow |[| and
% |]| in the title to toc.
{\protect\@title
{\smallerr% this macro is provided by the \pk{gmutils}
% package after the \pk{relsize} package.
\if@fshda\\[0.15em]\protect\@author
\if\relax\@date\relax\else, \fi
\else
\if\relax\@date\relax\else\\[0.15em]\fi
\fi
% The default is that all the included files have the same
% author(s). In this case we won't print the author(s) in
% the headings. Otherwise we wish to print them. The
% information which case are we in is brought by the
% |\if@fshda| switch defined in line \ref{@fshda}.
%
% If we wish to print the author's name
% (|\if@fshda|), then we'll print the date after the author, separated
% with a~comma. If we don't print the author, there still may be
% a~date to be printed. In such a~case we break the line, too, and
% print the date with no comma.
\protect\@date}}% end of |\incl@filedivtitle|'s brace (2nd
% or 3rd argument).
}% end of |\incl@filedivtitle|'s |\protected@xdef|.\par
% We |\protect| all the title components to avoid expanding
% |\footnotemark| hidden in |\thanks| during |\protected@xdef|
% (and to let it be executed during the typesetting, of course).
}% end of the comma-|\and|'s group.
\@xa\filediv\incl@filedivtitle
\@thanks
\g@relaxen\@author \g@relaxen\@title \g@relaxen\@date
\g@emptify\@thanks
}% end of |\InclMaketitle|.
% What I~make the default, is an assumption that all the
% multi-documented files have the same author(s).
% And with the account of the other possibility I~provide
% the below switch and declaration.
\newif\if@fshda
% \label{@fshda} (its name comes from \textit file\textit s
% \textit have \textit different \textit authors).
\newcommand*\PrintFilesAuthors{\@fshdatrue}
% And the counterpart, if you change your mind:
\newcommand*\SkipFilesAuthors{\@fshdafalse}
% \subdivision{The file's date and version information}
% Define |\filedate| and friends from info in the
% |\ProvidesPackage| etc.\ commands.
\def\GetFileInfo#1{%
\def\filename{#1}%
\def\gmu@tempb##1 ##2 ##3\relax##4\relax{%
\def\filedate{##1}%
\def\fileversion{##2}%
\def\fileinfo{##3}}%
\edef\gmu@tempa{\csname ver@#1\endcsname}%
\@xa\gmu@tempb\gmu@tempa\relax? ? \relax\relax}
% Since we may documentally input files that we don't load, as \docfm\
% e.g., let's define a~declaration to be put (in the comment layer)
% before the line(s) containing |\Provides...|. The |\FileInfo|
% command takes the stuff till the closing |]| and subsequent line
% end, extracts from it the info and writes it to the \file{.aux} and
% rescans the stuff. \eTeX\ provides a~special primitive for that
% action but we remain strictly \TeX nical and do it with writing to
% a~file and inputting that file.
\newcommand*\FileInfo{%
\bgroup
\gmd@ctallsetup
\bgroup% yes, we open two groups because we want to rescan tokens in
% `usual' catcodes. We cannot put \incs{gmd@ctallsetup} into the
% inner macro because when that will be executed, the
% \incs{inputlineno} will be too large (the last not the first line).
\let\do\@makeother
\do\ \do\{\do\}\do\^^M\do\\%
\gmd@fileinfo}
% \changes{v0.99d}{2007/04/26}{added}
\foone{%
\catcode`!\z@
\catcode`(\@ne
\catcode`)\tw@
\let\do\@makeother
\do\ % we make space `other' to keep it for scanning the code where
% it may be leading.
\do\{\do\}\do\^^M\do\\}%\CodeEscapeChar\!
(%
!def!gmd@fileinfo#1Provides#2{#3}#4[#5]#6^^M%
(!egroup% we close the group of changed catcodes, the catcodes
% of the arguments are set. And we are still in the group for
% \incs{gmd@ctallsetup}.
!gmd@writeFI(#2)(#3)(#5)%
!gmd@FIrescan(#1Provides#2{#3}#4[#5]#6)% this macro will close the group.
%\changes[\FileInfo]{v0.99m}{2008/08/09}{\cs{egroup} of the
% inner macro moved to the end to allow \cs{gmd@ctallsetup}. From the
% material passed to \cs{gmd@FIrescan} ending \hathat{M} stripped not
% to cause double labels.}
)% \CodeEscapeChar\\
)
\def\gmd@writeFI#1#2#3{%
% ^^A \typeout{@@@ write FI}\show\relax
\immediate\write\@auxout{%
\global\@nx\@namedef{%
ver@#2.\if P\@firstofmany#1\@@nil sty\else cls\fi}{#3}}}
\foone\obeylines{%
\def\gmd@FIrescan#1{%
% \changes{v0.99l}{2008/08/06}{\cs{scantokens} used instead of
% \cs{write} and \cs{@@input} which simplified the macro}
% ^^A \newwrite\gmd@docrescanfile
% ^^A \immediate\openout\gmd@docrescanfile=\jobname.docrescan\relax
{\newlinechar=`\^^M\scantokens{#1}}\egroup^^M}}
% And, for the case the input file doesn't contain |\Provides...|,
% a~macro for explicit providing the file
% info. It's written in analogy to |\ProvidesFile|, \pk{source ^^B
% 2${}_\epsilon$}, file L v1.1g, l.\,102.
\def\ProvideFileInfo#1{%
\begingroup
\catcode`\ 10 \catcode\endlinechar 10 %
\@makeother\/\@makeother\&%
\kernel@ifnextchar[{\gmd@providefii{#1}}{\gmd@providefii{#1}[]}%^^A]
}
% \changes{v0.98a}{06/09/06}{added}
% \DefIndex\gmd@providefii
\def\gmd@providefii#1[#2]{%\\
% (we \emph{don't} write the file info to \pk{.log})
\@xa\xdef\csname ver@#1\endcsname{#2}%
\endgroup}
% And a~self-reference abbreviation (intended for providing file info
% for the driver):
\def\ProvideSelfInfo{\ProvideFileInfo{\jobname.tex}}
% A~neat conventional statement used in \docfm's documentation e.g.,
% to be put in |\thanks| to the title or in a~footnote:
\newcommand*\filenote{This file has version number \fileversion{} dated \filedate{}.}
% And exactly as |\thanks|:
\newcommand*\thfileinfo{\thanks\filenote}
%
%
% \subdivision{Miscellanea}
%
% The main inputting macro, |\DocInput| has been provided. But there's
% another one in \docfm\ and it looks very reasonably:
% \gmhypertarget[IndexInput]{\cs{IndexInput}}. Let's make analogous one
% here:
\foone{\obeylines}%
{%
\def\IndexInput#1{% \changes{v0.98b}{06/09/07}{\cs{StoreMacro}ing
%% and \cs{RestoreMacro}ing
% \cs{code@delim} added}
\StoreMacro\code@delim%
\CodeDelim\^^Z%
\def\gmd@iihook{% \label{iihook}this hook is |\edef|ed!
\@nx^^M%
\code@delim\relax\@nx\let\@nx\EOFMark\relax}%^^A\@nx^^M
\DocInput{#1}\RestoreMacro\code@delim}%
}
% How does it work? We assume in the input file is no explicit
% \<char1>. This char is chosen as the code delimiter and will be put
% at the end of input. So, entire file contents will be scanned char
% by char as the code.
% \stanza
%
% \skiplines First I~tried to use \specialcomment{gmlonely} provided by
% the \pk{comment} package, but it caused errors.\endskiplines
%
% The below environment I~designed to be able to skip some repeating
% texts while documenting several packages of mine into one
% document. At the default settings it's just a~|\StraightEOL| group
% and in the |\skipgmlonely| declaration's scope it gobbles its
% contents.
%
\newenvironment{gmlonely}{\StraightEOL}{}
\newcommand\skipgmlonely[1][]{%
\def\gmu@tempa{%\DefIndex\gmd@skipgmltext
\def\gmd@skipgmltext{%
\g@emptify\gmd@skipgmltext
% ^^A \gmd@deeolize{
#1%^^A}
}}% not to count the lines of the substituting
% text but only of the text omitted
\gmu@tempa
\@xa\AtBegInput\@xa{\gmu@tempa}%
\renewenvironment{gmlonely}{%
\StraightEOL
\@fileswfalse% to forbid writing to \pk{.toc}, \pk{.idx} etc.
\setbox0=\vbox\bgroup}{\egroup\gmd@skipgmltext}}
% Sometimes in the commentary of this package, so maybe also others,
% I~need to say some char is of category 12 (`other sign'). This I'll
% mark just as \catother\ got by |\catother|.
%
\foone{\catcode`\_=8 }% we ensure the standard |\catcode| of |_|.
{
\newcommand*\catother{${}_{12}$}%\par
% Similarly, if we need to say some char is of category 13 (`active'),
% we'll write \catactive, got by |\catactive|
\newcommand*\catactive{${}_{13}$}%\par
% and a~letter, \catletter
\newcommand*\catletter{${}_{11}$}%.
}
%
% For the copyright note first I~used just \env{verse} but it requires
% marking the line ends with |\\| and indents its contents while
% I~prefer the copyright note to be flushed left. So
\newenvironment*{copyrnote}{%
\StraightEOL\everypar{\hangindent3em\relax\hangafter1 }%
\par\addvspace\medskipamount\parindent\z@\obeylines}{%
\@codeskipputgfalse\stanza}
%
% I~renew the \env{quotation} environment to make the fact of quoting
% visible.
\StoreEnvironment{quotation}
\def\gmd@quotationname{quotation}
\renewenvironment{quotation}{%
% \changes{v0.99j}{2008/08/03}{Improved behaviour of redefined
% \env{quotation} to be the original if used by another environment.}
% The first non-me user complained that \env{abstract} comes out in
% quotation marks. That is because \env{abstract} uses \env{quotation}
% internally. So we first check whether the current environment is
% \env{quotation} or something else.
\ifx\@currenvir\gmd@quotationname
\afterfi{\par``\ignorespaces}%
\else\afterfi{\storedcsname{quotation}}%
\fi}
{\ifx\@currenvir\gmd@quotationname
\afterfi{\ifhmode\unskip\fi''\par}%
\else\afterfi{\storedcsname{endquotation}}%
\fi}
% \chunkskip
% For some mysterious reasons |\noindent| doesn't work with the first
% (narrative) paragraph after the code so let's work it around:
\def\gmdnoindent{%
\ifvmode\leavevmode\hskip-\parindent\ignorespaces
\fi}% \incs{ignorespaces} is added to eat a~space inserted% by
% \incs{gmd@textEOL}. Without
% it it% also worked but it was a~bug: since% \incs{parindent} is
% a~dimen not% skip, \TeX\ looks forward and % expands macros to check
% whether% there is a~stretch or shrink part%% and therefore it
% gobbled the% \incs{gmd@textEOL}'s space.
%
% When a~verbatim text occurs in an inline comment, it's advisable to
% precede it with |%| if it begins a~not first line of such a~comment
% not to mistake it for a~part of code. Moreover, if such a~short verb
% breaks in its middle, it should break with the percent at the
% beginning of the new line. For this purpose provide
\newcommand*\inverb{%
% \changes{v0.99g}{2007/11/12}{added}
\@ifstar{%
\def\gmu@tempa{{\tt\xiipercent}}%
\@emptify\gmu@tempb% here and in the paralell points of the other
% case and \inverb|\nlpercent| I~considered an \inverb|\ifhmode|
% test but it's not possible to be in vertical mode while in an
% inline comment. If there happens vertical mode, the commentary
% begins to be `outline' (main text).
\gmd@inverb}%
{\@emptify\gmu@tempa
\def\gmu@tempb{\gmboxedspace}%
\gmd@inverb}}
\newcommand*\gmboxedspace{\hbox{\normalfont{ }}}
\newcommand*\gmd@nlperc[1][]{% \changes{v0.99n}{2008/08/22}{% added
% \cs{hbox}es in \cs{discretionary} to score \cs{hyphenpenalty} not
% \cs{exhyphenpenalty}}
\ifhmode\unskip\fi
\discretionary{\hbox{\gmu@tempa}}% (pre-break). I~always put
% a~\incs{hbox} here to make this discretionary score the
% \incs{hyphenpenalty} not \incs{exhyphenpenalty} (\TB\ p.~96) since
% the latter may be 10,000 in Polish typesetting.
{{\tt\xiipercent\gmboxedspace}}% (post-break)
{\gmu@tempb}% (no-break).
\penalty10000\hskip0sp\relax}
\newcommand*\gmd@inverb[1][]{%
\gmd@nlperc
\ifmmode\hbox\else\leavevmode\null\fi
\bgroup
\ttverbatim
\def\breakablevisspace{%
\discretionary{\visiblespace}{\xiipercent\gmboxedspace}{\visiblespace}}%
\def\breakbslash{%
\discretionary{}{\xiipercent\gmboxedspace\bslash}{\bslash}}%
\def\breaklbrace{%
\discretionary
{\xiilbrace\verbhyphen}%
{\xiipercent\gmboxedspace}%
{\xiilbrace}}%
\gm@verb@eol
%^^A \@ifstar {\@sverb@chbsl}
%^^A~ {\gmobeyspaces\frenchspacing\@sverb@chbsl}
\@sverb@chbsl% It's always with visible spaces.
}
\newcommand*\nlpercent{%
\@ifstar{\def\gmu@tempa{{\tt\xiipercent}}%
\@emptify\gmu@tempb
\gmd@nlperc}%
{\@emptify\gmu@tempa
\def\gmu@tempb{\gmboxedspace}%
\gmd@nlperc}}
\newcommand*\incs{% an inline \cs{cs}
% \changes{v0.99m}{008/08/07}{added}
\@ifstar{\def\gmu@tempa{{\tt\xiipercent}}%
\@emptify\gmu@tempb
\gmd@nlperc\cs}%
{\@emptify\gmu@tempa
\def\gmu@tempb{\gmboxedspace}%
\gmd@nlperc\cs}}
\def\inenv{\incs[]}% an in-line \cs{env}
% As you see, |\inverb| and |\nlpercent| insert a~discretionary that
% breaks to |%|
% at the beginning of the lower line. Without the break it's a~space
% (alas at its natural width i.e., not flexible) or, with the starred
% version, nothing. The starred version puts |%| also at the end of
% the upper line. Then |\inverb| starts sth. like
% |\verb*| but the breakables of it break to |%| in the
% lower line.
%
% \acro{TODO}: make the space flexible (most probably it requires using sth.\
% else than |\discretionary|).
%\stanza
%
% An optional hyphen for \CSs in the inline comment:
\@ifundefined{+}{}{\typeout{^^Jgmdoc.sty: redefining \bslash+.}}
\def\+{\discre{{\normalfont-}}{{\tt\xiipercent\gmboxedspace}}{}}
%
%
\providecommand*\ds{DocStrip}
% A~shorthand for |\CS|:
\pdef\CS{% \changes{v0.99n}{2008/08/30}{added}
\acro{CS}%
\@ifnextcat a{ }{}}% we put a~space if the next token is
% \catletter. It's the next best thing to checking whether the \CS
% consisting of letters is followed by a~space.
\pdef\CSs{\CS{}es\@ifnextcat a{ }{}}% for pluralis.
% \changes{v0.99n}{2008/08/30}{added}
\pdef\CSes{\CS{}es\@ifnextcat a{ }{}}% for pluralis.
% \changes{v0.99n}{2008/08/30}{added}
% \gmhypertarget[CDAnd]{Finally,} a~couple of macros for documenting
% files playing with |%|'s
% catcode(s). Instead of |%| I~used |&|. They may be at the end
% because they're used in the commented thread i.e.\ after package's
% |\usepackage|.
\newcommand*\CDAnd{\CodeDelim\&}
\newcommand*\CDPerc{\CodeDelim*\%}
% And for documenting in general:
%
% A~general sectioning command because I~foresee a~possibility of
% typesetting the same file once as independent document and another
% time as a~part of bigger whole.
% \Define\division
\let\division=\section
% \Define\subdivision
\let\subdivision=\subsection
% \Define\subsubdivision
\let\subsubdivision=\subsubsection
%
%
% \stanza
% To kill a~tiny little bug in \file{doc.dtx} (in line 3299 |\gmu@tempb| and
% |\gmu@tempc| are written plain not verbatim):
\newcounter{gmd@mc}
% Note it is after the \env{macrocode} group
\def\gmd@mchook{\stepcounter{gmd@mc}% \label{gmd@mchook}
\gmd@mcdiag
\ifcsname gmd@mchook\the\c@gmd@mc\endcsname
\afterfi{\csname gmd@mchook\the\c@gmd@mc\endcsname}%
\fi}
\long\def\AfterMacrocode#1#2{\@namedef{gmd@mchook#1}{#2}}
% What have I~done? I~declare a~new counter and employ it to count the
% \env{macrocode(*)}s (and \env{oldmc(*)}s too, in fact) and attach
% a~hook to (after) the end of every such environment. That lets us to
% put some stuff pretty far inside the compiled file (for the buggie
% in \file{doc.dtx}, to redefine |\gmu@tempb/c|).
% One more detail to expalin and define: the |\gmd@mcdiag| macro may
% be defined to type out a~diagnostic message (the
% \env{macrocode(*)}'s number, code line number and input line number).
\@emptify\gmd@mcdiag
\def\mcdiagOn{\def\gmd@mcdiag{%
\typeout{^^J\bslash end{\@currenvir} No.\the\c@gmd@mc
\space\on@line, cln.\the\c@codelinenum.}}}
\def\mcdiagOff{\@emptify\gmd@mcdiag}
%
% An environment to display the meaning of macro parameters: its
% items are automatically numbered as |#1|, |#2| etc.
\newenvironment*{enumargs}[1][1]%
% \changes{v0.99n}{2008/08/21}{developed for the case of inline
% comment}
% \changes{v0.99o}{2008/09/04}{added the optional argument which is
% the number of hashes (1 by default or 2 or 4)}
% \changes{v0.99p}{2008/10/4}{added optional arguments' handling}
{\if@aftercode\edef\gmu@tempa{\the\leftskip}%
\edef\gmu@tempb{\the\hangindent}\fi
\enumerate
\if@aftercode
\leftskip=\glueexpr\gmu@tempa+\gmu@tempb\relax
\fi
\@namedef{label\@enumctr}{%
\env{\if@aftercode\code@delim\space\fi
\gmd@ea@bwrap
\#\ifcase#1\relax\or\or\#\or\or\#\#\#\fi
\csname the\@enumctr\endcsname
\gmd@ea@ewrap}}%
\let\mand\item
\provide\gmd@ea@wraps{%
\emptify\gmd@ea@ewrap
\emptify\gmd@ea@bwrap}%
\gmd@ea@wraps
\def\opt{%
\def\gmd@ea@bwrap{[}\def\gmd@ea@ewrap{]}%
\item
\gmd@ea@wraps}%
}
{\endenumerate}
% The starred version is intended for lists of arguments some of which
% are optional: to align them in line.
\newenvironment*{enumargs*}{%
\def\gmd@ea@wraps{%
\def\gmd@ea@bwrap{ }\def\gmd@ea@ewrap{ }}%
\enumargs}{\endenumargs}
%
% \subdivision[\docfm-compatibility]{\gmhypertarget[doccompa]{\docfm-compatibility}}
%
% \gmhypertarget[doc-likeness]{My} \TeX\ Guru recommended me to write
% hyperlinking for \docfm. The suggestion came out when writing of
% \gmdoc\ was at such a~stage that I~thought it to be much easier to write
% a~couple of |\let|s to make \gmdoc\ able to typeset
% sources written for \docfm\ than to write a~new package that adds
% hyperlinking to \docfm. So\dots
%
% \stanza
% The \docfm\ package makes |%| an ignored char. Here the |%| delimits
% the code and therefore has to be `other'. But only the first one
% after the code. The others we may re|\catcode| to be ignored and we
% do it indeed in line \ref{ignorePercent}.
% \stanza
%
% At the very beginning of a~\docfm-prepared file we meet a~nice
% command \cs{Char\-act\-er\-Tab\-le}. My \TeX\ Guru says it's a~bit old
% fashioned these days so let's just make it notify the user:
\def\CharacterTable{\begingroup
\@makeother\{\@makeother\}%
\Character@Table}
\foone{%
\catcode`\[=1 \catcode`\]=2 %
\@makeother\{\@makeother\}}%
[
\def\Character@Table#1{#2}[\endgroup
\message[^^J^^J gmdoc.sty package:^^J
==== The input file contains the \bslash CharacterTable.^^J
==== If you really need to check the correctness of the chars,^^J
==== please notify the author of gmdoc.sty at the email address^^J
==== given in the legal notice in gmdoc.sty.^^J^^J]%
%^^A~\gmd@countlines[#1]\gmd@countlines[#2]
]]
% Similarly as \docfm, \gmdoc\ provides \env{macrocode}, \env{macro}
% and \env{environment} environments. Unlike in \docfm,
% |\end{macrocode}| \emph{does not} require to be preceded with any
% particular number of spaces. Unlike in \docfm, it \emph{is not}
% a~kind of \env{verbatim}, however, which means the code and
% narration layers remains in force inside it which means that any
% text after the first |%| in a~line will be processed as narration
% (and its control sequences will be executed). For a~discussion of
% a~possible workaround see line \ref{AVerySpecialMacro}.
%
% \stanza
% Let us now look over other original \docfm's control sequences and
% let's `domesticate' them if they are not yet.
%
% The \TextUsage\DescribeMacro\ and \TextUsage\DescribeEnv\ commands
% seem to correspond with my |\TextUsage| macro in its plain and starred
% version respectively except they don't typeset their arguments in the text
% i.e., they do two things of the three. So let's |\def| them to do these
% two things in this package, too:
\outer\def\DescribeMacro{%
\begingroup\MakePrivateLetters
\gmd@ifonetoken\Describe@Macro\Describe@Env}
% Note that if the argument to |\DescribeMacro| is not
% a~(possibly starred) control sequence, then as an environment's name
% shall it be processed \emph{except} the \cs{Mak\-ePriv\-at\-e\-Oth\-ers}
% re|\catcode|ing shall not be done to it.
\outer\def\DescribeEnv{%
\begingroup\MakePrivateOthers\Describe@Env}
% Actually, I've used the |\Describe...| commands myself a~few times, so
% let's |\def| a~common command with a~starred
% version:
\outer\def\Describe{% It doesn't typeset its argument in the point of
% occurrence.
\begingroup\MakePrivateLetters
\@ifstarl{\MakePrivateOthers\Describe@Env}{\Describe@Macro}}
% The below two definitions are adjusted \*s of |\Text@UsgMacro| and
% |\Text@UsgEnvir|.
% \DefIndex\Describe@Macro
\long\def\Describe@Macro#1{%
\endgroup
\strut\Text@Marginize#1%
\@usgentryze#1% we declare kind of formatting the entry
\text@indexmacro#1\ignorespaces}
% \DefIndex\Describe@Env
\def\Describe@Env#1{%
\endgroup
\strut\Text@Marginize{#1}%
\@usgentryze{#1}% we declare the `usage' kind of
% formatting the entry and index the sequence |#1|.
\text@indexenvir{#1}\ignorespaces}
% Note that here the environments' names are typeset in |\tt| font
% just like the macros', \emph{unlike} in \docfm.
%
% \stanza
% My understanding of `minimality' includes avoiding too much freedom
% as causing chaos not beauty. That's the philosophical and \ae^^B
% sthetic reason why I~don't provide \TextUsage\MacroFont. In
% my opinion there's a~noble tradition of typesetting the \TeX\ code
% in |\tt| font nad this tradition sustained should be. If one wants
% to change the tradition, let \himher\ redefine |\tt|, in \TeX\ it's
% no problem. I~suppose |\MacroFont| is not used explicitly, and that
% it's (re)defined at most, but just in case let's |\let|:
\let\MacroFont\tt
%
% \stanza
% We have provided \TextUsage\CodeIndent\ in line
% \ref{CodeIndent}. And it corresponds with \docfm's
% \Describe\MacroIndent\cs{Mac\-ro\-In\-dent} so
% \Define\MacroIndent
\let\MacroIndent\CodeIndent
% And similarly the other skips: \Define\MacrocodeTopsep
\let\MacrocodeTopsep\CodeTopsep
% Note that \TextUsage\MacroTopsep\ is defined in \gmdoc\ and has
% the same r\ocircum le as in \docfm.
% \Define\SpecialEscapechar
\let\SpecialEscapechar\CodeEscapeChar
% \TextUsage\theCodelineNo\ is not used in
% \pk{gmdoc}. Instead of it there is \TextUsage\LineNumFont\
% declaration and a~possibility to redefine |\thecodelinenum| as for
% all the counters. Here the |\LineNumFont| is used two different
% ways, to set the benchmark width for a~linenumber among others, so it's not
% appropriate to put two things into one macro. Thus let's give the
% user a~notice if \heshe\ defined this macro:
%
% Because of possible localness of the definitions it seems to be
% better to add a~check at the end of each |\DocInput| or
% |\IndexInput|.
\AtEndInput{\@ifundefined{theCodelineNo}{}{\PackageInfo{gmdoc}{The
\string\theCodelineNo\space macro has no effect here, please use
\string\LineNumFont\space for setting the font and/or
\string\thecodelinenum\space to set the number format.}}}
% I~hope this lack will not cause big trouble.
%
% \stanza
% For further notifications let's define a~shorthand:
\def\noeffect@info#1{\@ifundefined{#1}{}{\PackageInfo{gmdoc}{^^J%
The \bslash#1 macro is not supported by this package^^J
and therefore has no effect but this notification.^^J
If you think it should have, please contact the maintainer^^J
indicated in the package's legal note.^^J}}}
% The four macros formatting the macro and environment names, namely
% \possfil\TextUsage\PrintDescribeMacro,
% \possfil\TextUsage\PrintMacroName,
% \possfil\TextUsage\PrintDescribeEnv\ and
% \possfil\TextUsage\PrintEnvName\ are not
% supported by \pk{gmdoc}. They seem to me to be too internal to take
% care of them. Note that in the name of (\ae sthetical) minimality
% and (my) convenience I~deprive you of easy knobs to set strange
% formats for verbatim bits: I~think they are not advisable.
%
% Let us just notify the user.
\AtEndInput{%
\noeffect@info{PrintDescribeMacro}%
\noeffect@info{PrintMacroName}%
\noeffect@info{PrintDescribeEnv}%
\noeffect@info{PrintEnvName}}
% The \TextUsage\CodelineNumbered\ declaration of \docfm\ seems to be
% equivalent to our |noindex| option with the |linesnotnum| option set
% off so let's define it such a~way.
\def\CodelineNumbered{\AtBeginDocument{\gag@index}}
\@onlypreamble\CodelineNumbered
% Note that if the |linesnotnum| option is in force, this declaration
% shall not revert its effect.
%
% I~assume that if one wishes to use \docfm's interface then \heshe'll
% not use \gmdoc's options but just the default.
%
% \stanza
% The |\CodelineIndex| and |\PageIndex| declarations
% correspond with the \pk{gmdoc}'s default and the |pageindex| option
% respectively.
% Therefore let's |\let|
\let\CodelineIndex\@pageindexfalse
\@onlypreamble\CodelineIndex
\let\PageIndex\@pageindextrue
% \label{PageIndex}
\@onlypreamble\PageIndex
% The next two declarations I~find useful and
% smart:
\def\DisableCrossrefs{\@bsphack\gag@index\@esphack}
\def\EnableCrossrefs{\@bsphack\ungag@index
\def\DisableCrossrefs{\@bsphack\@esphack}\@esphack}
% The latter definition is made due to the footnote 6 on p.\,8 of the
% Frank Mittelbach's \docfm's documentation and both of them are
% copies of lines 302--304 of it modulo |\|(|un|)|gag@index|.
%
% \stanza
% The subsequent few lines I~copy almost verbatim ;-) from the lines
% 611--620.
\newcommand*\AlsoImplementation{\@bsphack%
\long\def\StopEventually##1{\gdef\Finale{##1}}% we define
% \cs{Fi\+n\+a\+le} just to expand to the argument of |\StopEventually| not
% to to add anything to the end input hook because |\Finale| should
% only be executed if entire document is typeset.\nostanza\par
%
% \hangindent\verbatimhangindent \hangafter1\relax
% \gmdnoindent \leftskip\CodeIndent
% |%\init@checksum| is obsolete in
% \pk{gmdoc} at this point: the \env{CheckSum} counter is reset just at
% the beginning of (each of virtually numerous) input(s).
% \nostanza\par
\@esphack}
\AlsoImplementation
% \begin{quotation} When the user places an |\OnlyDescription|
% declaration in the driver file the document should only be typeset
% up to |\StopEventually|. We therefore have to redefine this
% macro.\end{quotation}
\def\OnlyDescription{\@bsphack\long\def\StopEventually##1{%
% \begin{quotation} In this case the argument of |\StopEventually|
% should be set and afterwards \TeX\ should stop reading from
% this file. Therefore we finish this macro with\end{quotation}
##1\endinput}\@esphack}
% \begin{quotation} If no |\StopEventually| command is given we
% silently ignore a~|\Finale| issued.\end{quotation}
\@relaxen\Finale
% The \TextUsage\meta\ macro is so beautifully crafted in \docfm\ that
% I~couldn't resist copying it into \pk{gmutils}. It's also available
% in Knuthian (\TeXbook\ format's) disguise
% \Describe*{\<...>}|\<|\<the argument>|>|.
% \stanza
% The checksum mechanism is provided and developed for a~slightly
% different purpose.
% \stanza
% Most of \docfm's indexing commands have already been `almost
% defined' in \pk{gmdoc}:
\let\SpecialMainIndex=\DefIndex
% \DoNotIndex\endcsname*
\def\SpecialMainEnvIndex{\csname CodeDefIndex\endcsname*}% we don't
% type |\DefIndex| explicitly here because it's |\outer|, remember?
% \Define\SpecialIndex \Define\SpecialUsageIndex
% \Define\SpecialEnvIndex \Define\SortIndex
\let\SpecialIndex=\CodeCommonIndex
\let\SpecialUsageIndex=\TextUsgIndex
\def\SpecialEnvIndex{\csname TextUsgIndex\endcsname*}
\def\SortIndex#1#2{\index{#1\actualchar#2}}
% \begin{quotation}All these macros are usually used by other macros;
% you will need them only in an emergency.\end{quotation}
%
% Therefore I~made the assumption(s) that `Main' indexing macros are
% used in my `Code' context and the 'Usage' ones in my `Text' context.
%
% \stanza
% Frank Mittelbach in \docfm\ provides the \TextUsage\verbatimchar\
% macro to (re)define the |\verb(*)|'s delimiter for the index
% entries. The \gmdoc\ package uses the same macro and its default definition
% is |{&}|. When you use \docfm\ you
% may have to redefine |\verbatimchar| if you use (and index) the |\+|
% control sequence. \pk{gmdoc} does a~check for the analogous situation
% (i.e., for processing |\&|) and
% if it occurs it takes |$| as the |\verb*|'s delimiter. So strange
% delimiters are chosen deliberately to allow any `other' chars in the
% environments' names. If this would cause problems, please
% notify me and we'll think of adjustments.
%
\def\verbatimchar{&}
% \changes{v0.98c}{06/9/10}{put into all indexing macros for the
% accordance of the `macro' and the `environment' index entries; the
% \cs[]{\$} sign set as its alternative}
% One more a~very neat macro provided by \docfm. I~copy it
% verbatim and put into \pk{gmutils}, too. (|\DeclareRobustCommand|
% doesn't issue an error if its argument has been defined, it only
% informs about redefining.)
\pdef\*{\leavevmode\lower.8ex\hbox{$\,\widetilde{\ }\,$}}
% \changes{v0.98}{06/09/05}{made robust.}
% \TextUsage\IndexPrologue\ is defined in line
% \ref{IndexPrologue}. And other \docfm\ index commands too.
\@ifundefined{main}{}{\let\DefEntry=\main}
\@ifundefined{usage}{}{\let\UsgEntry=\usage}
% About how the \ds\ directives are supported by \pk{gmdoc},
% see section \gmiflink[docstrip]{The \ds\dots}.
% This support is not \emph{that} sophisticated as in \docfm, among others, it
% doesn't count the modules' nesting. Therefore if we dont want an
% error while \pk{gmdoc}umenting \docfm-prepared files, better let's
% define \docfm's counter for the modules' depths.
\newcounter{StandardModuleDepth}
% \stanza
% For now let's just mark the macro for further development
% \Define*\DocstyleParms
\noeffect@info{DocstyleParms}
% For possible further development or to notify the user once and
% forever:
% \Define\DontCheckModules \Define\CheckModules
\@emptify\DontCheckModules \noeffect@info{DontCheckModules}
\@emptify\CheckModules \noeffect@info{CheckModules}
% The \TextMarginize\Module|\Module| macro \emph{is} provided exactly
% as in \docfm.
% \Define\AltMacroFont
\@emptify\AltMacroFont \noeffect@info{AltMacroFont}
% \begin{quotation} And finally the most important bit: we change the
% |\catcode| of |%| so that it is ignored (which is how we are able to
% produce this \pk{doc}ument!). We provide two commands to do the
% actual switching.\end{quotation}
\def\MakePercentIgnore{\catcode`\%9\relax}
\def\MakePercentComment{\catcode`\%14\relax}
%
%
% \subdivision{\gmdoc ing \file{doc.dtx}}
%
% The author(s) of \docfm\ suggest(s): \begin{quotation}For examples
% of the use of most---if not all---of the features described above
% consult the \file{doc.dtx} source itself.\end{quotation}
% Therefore I~hope that after \file{doc.dtx} has been \gmdoc-ed, one
% can say \gmdoc\ is \docfm-compatible ``at most---if not at all''.
%
% \TeX ing the original \docfm\ with my humble\footnote{What ^^B(
% a~\emph{false} modesty! ;-)} package was a~challenge and
% a~milestone experience in my \TeX\ life.
%
% \stanza
% One of minor errors was caused by my understanding of a~`shortverb'
% char: due to \pk{gmverb}, in the math mode an active `shortverb' char
% expands to itself's `other' version thanks to |\string| (It's done
% with \verb+|+ in mind). \docfm's
% concept is different, there a~`shortverb' char should in the math
% mode work as shortverb. So let it be as they wish: \pk{gmverb}
% provides |\OldMakeShortVerb| and the oldstyle input commands change
% the inner macros so that also |\MakeShortVerb| works as in \docfm
% (cf.\ line \ref{oldmkshvrrb}).
%
% We also redefine the \env{macro} environment to make it mark the
% first code line as the point of defining of its argument, because
% \pk{doc.dtx} uses this environment also for implicit definitions.
%
% \changes[\OldMakeShortVerb]{v0.98a}{06/09/05}{and pals moved to
% \pk{gmverb}.}
%\Define\OldDocInput
% \changes{v0.98}{06/09/05}{\cs{@makeother\protect\bslash_} added}
% \changes{v0.98a}{06/09/05}{\cs{AtBegInput} changed into
% \cs{AtBegInputOnce}.}
% \changes{v0.98b}{06/09/07}{enrichments of the \env{macrocode(*)}
% definitions moved to the default definitions of these envs.}
\def\OldDocInput{% \changes{v0.99g}{2007/11/11}{obsolete redefinition
% of the \env{macro} environment removed}
\AtBegInputOnce{\StraightEOL
\let\@MakeShortVerb=\old@MakeShortVerb
% \label{oldmkshvrrb}
\OldMacrocodes}%
\bgroup\@makeother\_% it's to allow |_| in the filenames. The next
% macro will close the group.
\Doc@Input}
% We don't swith the |@codeskipput| switch neither we check it because
% in `old' world there's nothing to switch this switch in the
% narration layer.
%
% \stanza
% I~had a~hot and wild \TeX\ all the night nad what a~bliss when
% the `Succesfully formated 67 page(s)' message appeared.
%
% My package needed fixing some bugs and adding some compatibility
% adjustments (listed in the previous section) and the original
% \pk{doc.dtx} source file needed a~few adjustments too because some
% crucial differences came out. I'd like to write a~word about them now.
%
% \stanza
% The first but not least is that the author(s) of \docfm\ give the
% \CS marking commands non-macro arguments sometimes, e.g.,
% |\DescribeMacro{StandardModuleDepth}|. Therefore we should launch
% the \emph{starred} versions of corresponding \gmdoc\ commands. This
% means the \docfm-like commands will not look for the \CS's occurrence
% in the code but will mark the first codeline met.
% \stanza
%
% Another crucial difference is that in \gmdoc\ the narrative and the
% code layers are separated with only the code delimiter and
% therefore may be much more mixed than in \docfm. among others, the \env{macro}
% environment is \emph{not} a~typical \env{verbatim} like: the texts
% commented out within \env{macrocode} are considered a~normal
% commentary i.e., not verbatim. Therefore some macros `commented out'
% to be shown verbatim as an example source must have been `additionally'
% verbatimized for \gmdoc\ with the shortverb chars e.g. You may also
% change the code delimiter for a~while, e.g., the line
% \CodeDelim\.
% \AVerySpecialMacro % delete the first % when.\unskip|..|\par\CDPerc
%
% \gmdnoindent\label{AVerySpecialMacro}
% was got with
% \begin{verbatim}
%\CodeDelim\.
%% \AVerySpecialMacro % delete the first % when.\unskip|..|\CDPerc
%\end{verbatim}
%
% \leftskip0sp\relax
% One more difference is that my shortverb chars expand to their
% \catother\ versions in the math mode while in \docfm\ remain
% shortverb, so I~added a~declaration |\OldMakeShortVerb| etc.
%
% \stanza
% Moreover, it's \TeX ing \docfm\ what inspired adding the
% |\StraightEOL| and |\QueerEOL| declarations.
%
%
%
% \division{Polishing, development and bugs}
%
% \begin*{bulletpars}
% \everypar{$\bullet$ }
% \label{Tasks}|\MakePrivateLetters| theoretically may interfere
% with |\active|ating some chars to allow linebreaks. But making
% a~space or an opening brace a~letter seems so perverse that we may
% feel safe not to take account of such a~possibility.
%
% When |countalllines*| option is enabled, the comment lines
% that don't produce any printed output result with a~(blank) line too
% because there's put a~hypertarget at the beginning of them. But for
% now let's assume this option is for draft versions so hasn't be
% perfect.
%
% Marcin Woli\nacute ski suggests to add the marginpar clauses for the
% \acro{AMS} classes as we did for the standard ones in the lines
% \ref{mparclause1}--\ref{mparclause2}. Most probably I~can do it on
% request when I~only know the classes' names and their `marginpar
% status'.
%
% When the |countalllines*| option is in force, some |\list|
% environments shall raise the `missing |\item|' error if you don't
% put the first |\item| in the same line as
% |\begin{|\<environment>|}| because the (comment-) line number is
% printed.^^A~\end{}
%
% I'm prone to make the control sequences hyperlinks to the(ir)
% `definition' occurrences. It doesn't seem to be a~big work compared
% with what has been done so far.
%
% Is |\RecordChanges| really necessary these days? Shouldn't be the
% |\makeglossary| command rather executed by default?\footnote{It's ^^B
% understandable that ten years earlier writing things out to the files ^^B
% remarkably decelerated \TeX, but nowadays it does not in most ^^B
% cases. That's why \cs{makeindex} is launched by default in \gmdoc.}
%
% Do you use |\listoftables| and/or |\listoffigures| in your
% documentations? If so, I~should `\acro{EOL}-straighten' them like
% |\tableofcontents|, I~suppose (cf.\ line \ref{straighttoc}).
%
% Some lines of non-printing stuff such as |\Define...| and
% |\changes| connecting the narration with the code resulted with
% unexpected large vertical space. Adding a~fully blank line between
% the printed narration text and not printed stuff helped.
%
% Specifying |codespacesgrey, codespacesblank| results in typesetting
% all the spaces grey including the leading ones.
%
%
% About the \ds\ \gmiflink{verbatim mode directive} see above.
%
% \end{bulletpars}
% \ \par\vspace*{-\baselineskip}
%
% \def\EOFMark{\<eof>}
% \division{(No) \EOFMark}
%
% Until version 0.99i
% a~file that is |\DocInput| had to be ended with a~comment line with
% an |\EOF| or |\NoEOF| \CS that
% suppressed the end-of-file character to make input end
% properly. Since version 0.99i however the proper ending of input is
% acheved with |\everyeof| and therefore |\EOF| and |\NoEOF| become
% a~bit obsolete.
%
% If the user doesn't wish the documentation to be ended by
% `\EOFMark', \heshe\ should redefine the |\EOFMark| \CS or end
% the file with a~comment ending with |\NoEOF| macro defined
% below\footnote{Thanks to Bernd Raichle at Bacho\TeX\ 2006 ^^B Pearl
% Session where he presented \cs{input}ing a~file inside ^^B
% \cs{edef}.}:
%
\foone{\catcode`\^^M\active }{%
\def\@NoEOF#1^^M{%
\@relaxen\EOFMark\endinput}%
\def\@EOF#1^^M{\endinput}}
\def\NoEOF{\QueerEOL\@NoEOF}
\def\EOF{\QueerEOL\@EOF}
% \changes{v0.98}{06/09/05}{extended to add it the \cs{endinput} effect}
% \changes{v0.98l}{06/10/27}{divided in two macros first of which
% makes queer \acro{EOL} and the latter gobbles the stuff till the \acro{EOL} to
% suppress possible \cs{endinput} when used in \cs{StopEventually}}
% \label{NoEOF} As you probably see, \cs{(No)EOF} have the `immediate'
% |\endinput| effect: the file ends even in the middle of a~line, the
% stuff after \cs{(No)EOF} will be gobbled unlike with a~bare
% |\endinput|.
\endinput
%^^A~\docstrip
%</package>
%\PrintChanges \PrintIndex
%
%
%^^A~place for general changes:
% \ChangesGeneral
%
% \changes{v0.98a}{06/09/06}{\cs{AtBeginput}, \cs{AtEndinput},
% \cs{AtBeginputOnce} renamed to \cs{AtBegInput}, \cs{AtEndInput}
% \cs{AtBegInputOnce} respect.}
%
% \changes{v0.98c}{06/9/9}{making \cs{Define} and \cs{CodeUsage}
% markers to add up; bug fixes in indexing macros and change of
% concept of \cs{Code(Define$\quotechar|$Usage)*}: hence they serve not only for
% environments but also for implicit def/use of macros}
%
% \changes{v0.98d}{06/9/11}{mostly editing the narrative, marking
% special index entries etc.; queering \cs{char1} and \cs{char2}
% repeated \cs{AtBegInput}}
%
% \changes{v0.98e}{06/09/24}{Some macros moved to \pk{gmutils}:
% \cs{cs} and pals and one-chars with queer \cs{catcode}s}
%
% \changes{v0.98f}{06/09/27}{A~bug fixed: \cs{Code@MarginizeMacro} set
% to define a~\CS and the corresponding test set to check if it's
% undefined. In all three such definitions and resets after the use,
% \cs{def} is changed to \cs{(g)let}}
%
% \changes{v0.98g}{06/10/10}{among others, \cs{discretionary}s for breaking
% a~\CS to percent at the beginning of the lower line. Moreover, fixing
% a~bug/feature that leaves the code leftskip in the narration when an
% inline comment is followed by another codeline (w.o.\ explicit
% \cs{par}). And lots of finishing touches to the text. A~special font
% for the marginpar \CSs among others}
%
% \changes{v0.98l}{06/10/26}{in \cs{AtBegInputOnce} an auxiliary macro
% for each use substituted with one macro added at begin input. In
% \cs{gmd@evpaddonce} a~counter substituted with a~numeric
% macro. \cs{@ifQueerEOL} made polite i.e., a~two-argument not
% expanding to an open \cs{if...}}
%
% \changes{v0.99}{06/11/24}{\env{oldmc(*)} implemented that's
% (hope)fully compatible with \docfm's \env{macrocode(*)}. Moreover,
% a~declaration letting \gmdoc's \env{macrocode} to \env{oldmc(*)}}
%
% \changes{v0.99a}{06/11/30}{The \ds\ directives implemented fully
% automatic (no more need of \cs{doctrip(s)} declarations). Moreover,
% some minor changes due to \TeX ing The Source.}
%
% \changes{v0.99b}{2007/01/01}{Thanks to the \cs{edverbs} declaration
% in the class, displayed shortverbs simplified; Emacs mode changed to
% doctex. Author's true name more exposed}
%
% \changes{v0.99c}{2007/03/02}{A~bug fixed in \cs{DocInput} and all
% \cs{expandafter}s changed to \cs{@xa} and \cs{noexpand}s to
% \cs{@nx}}
%
% \changes{v0.99c}{2007/03/12}{The \TeX-related logos now are declared
% with \cs{DeclareLogo} provided in \pk{gmutils}}
%
% \changes{v0.99d}{2007/04/13}{\cs{afterfi} \& pals made two-argument}
%
% \changes{v0.99d}{2007/04/24}{\cs{@namelet} renamed to \cs{n@melet}
% to solve a~conflict with the \pk{beamer} class (in \pk{gmutils} at
% first)}
%
% \changes{v0.99e}{2007/04/29}{a~bug fixed in \cs{DocInput} and
% \cs{IndexInput}}
%
% \changes{v0.99g}{2007/11/12}{The bundle goes \XeTeX. The
% \TeX-related logos now are moved to \pk{gmutils}. \hathat{A}
% becomes more comment-like thanks to re\cs{catcode}'ing. Automatic
% detection of definitions implemented}
%
% \changes{v0.99h}{2007/11/16}{Fixed behaviour of sectioning commands
% (optional two heading skip check) of \pk{mwcls/gmutils} and
% respective macro added in \pk{gmdocc}. I~made a~\file{tds} archive}
%
% \changes{v0.99i}{2008/08/03}{A~“feature not bug” fix: thanks to
% \cs{everyeof} the \cs{(No)EOF} is now not necessary at the end of
% \cs{DocInput} file.}
%
% \changes{v0.99m}{2008/08/08}{Counting of all lines developed (the
% \env{countalllines} package option), now it
% uses \cs{inputlineno}}
%
% \changes{v0.99n}{2008/08/24}{Inline comments' alignment developed}
%
% \Finale
% (For my GNU Emacs:)
%%% Local Variables:
%%% mode: doctex
%%% End: